Advertise on podcast: English language Visionary Marketing Podcasts

Mike Hadjadj, a retail and SaaS expert, has just published a book (in French) called SaaSpocalypse. Every year, he supports two to three hundred start-ups as a mentor, advisor and investor, mostly in the retail sector. In his own words, he isn’t predicting the end of software vendors, but he does see their value being hit hard by the changes AI is bringing. I interviewed him to understand what this shake-up means for vendors, their customers and the whole SaaS ecosystem. SaaSpocalypse Now! And The Value Drain Mike Hadjadj presenting his book SaaSpocalypse (in French) at the Visionary Marketing offices in October 2026 TL;DR With SaaSpocalypse, Mike Hadjadj isn’t predicting the end of software vendors. He’s saying the price of software is now negotiable. French e-commerce giant La Redoute rebuilt a €360,000-a-year tool in 48 hours using AI. Three factors can shield a SaaS vendor, namely urgency, value and stickiness. Big vendors are not spared. Some retailers have halved their Salesforce budget, or even cut it to a third. Gartner is looking in the rear-view mirror, while Mike is watching the 2027 contract renewals. His advice to vendors is simple. Talk to your customers now and ask them why they pay. SaaSpocalypse Now! or How to Build a SaaS Without Coding M.H. The book started with a chat at my physio’s. I’d had an accident and needed rehab. He’s a bit of a geek and told me he was setting up a new business line without buying any software or paying for a single SaaS subscription. Using Base44, a tool owned by Wix, he’d rebuilt a small CRM, a training tool and a small accounting and invoicing package himself. That intrigued me. I run La Retail Tech, a community that has been going for four and a half years, with lots of content and around 3,000 members. Our community platform was grossly outdated. It had stopped evolving and every small request was sitting in a support ticket for three weeks. We Dreamed It, and AI Delivered I don’t sleep much, so I gave it a go and I spent a few hours talking to an LLM, asking it to rework parts of the site and give it a modern look. I ended up with a front end, which I showed my team after the weekend. They recognised the site they’d been dreaming of. We had almost launched a tender, complete with requirements definition and technical provisos. That meant bickering over compromises with the team, because not everything is possible. With AI, everything is crystal clear. I talk and hey presto! It’s done. SaaSpocalypse Now! Mike Hadjadj during a demo for his community, La Retail Tech. Image by Optimalways I’ve kept going since the book came out. The front end is done and I’m now tackling the back end. The agent kept misunderstanding me, so I found another agent that handled my agent, which I talk to on WhatsApp. It knows my whole environment, runs the tests and writes very detailed prompts that I just copy and paste. I can’t code, and neither can my physio. So the question isn’t whether people can do it. It’s about who will still pay, for what, and where the value sits. That’s one of the book’s main stances. A French E-commerce Giant Rebuilt a €360,000-a-Year Tool in 48 Hours M.H. The other trigger was meeting French e-commerce giant La Redoute at an event. They used a visual product discovery tool, created by an Israeli start-up. The tool worked very well. Other French retailers used it too, such as Monoprix and Castorama, but you could tell it was nearing the end of its life. They were all looking for a replacement and asking me if I knew of one, before launching a tender. La Redoute put two people on it for 48 hours, with AI. They studied how the tool worked, rebuilt it, tested it and released it. The copy is better than the original, because they could tune it to their own environment, something that a software vendor seldom allows. La Redoute was spending €360,000 a year on the tool, now they don’t. SaaSpocalypse Companies of All Sizes Are Following Suit Digging deeper, I found out that companies of all sizes, including large ones with the right skills, are already going after commoditised tools that aren’t core to the business. If the tool falls over, it’s no big deal. The savings are huge. When I talk about value destruction, it’s because the rebuild cost nothing like the old price. Two days, two people, a bit of maintenance, updates, testing and tokens. There’s simply no comparison. A small part of the value shifts to in-house skills, and some is bought externally, at a very different price. Most of it is destroyed. There’s also a whole industry around SaaS spending. Agencies, consultancies, integrators and, above all, start-ups, which means funds and fundraising. If the pie shrinks, all of that ecosystem will be hit. Urgency, Value or Stickiness, What Protects a Software Vendor? M.H. I’ve been supporting start-ups for a very long time, as a mentor, an advisor, in incubators and accelerators, and as an investor too. Until now, I judged them on unwritten criteria, to decide whether I believed in them, wanted to back them or even invest. How innovative they were, how good the team was. That has changed, and I’ve come up with a new three-part test. The first criterion is urgency. The need has to be pressing, driven for instance by regulation or competition. The second is value, which must be blatant, immediate and measurable. The third is new to me, and it’s stickiness. This is what will refrain end customers or competitors from copying your product in a few minutes. SaaSpocalypse, an Innovator “Disrupted” by Its Own Customers In the book, I cite Smartway, a pioneer in the fight against food waste. Its software tracked the best-before dates of perishable products in stores, to decide when to pull an item from the shelf and what discount would guarantee a sale. A salad or a yoghurt might go at 22% off one day and 40% off the next. It first sold well to independent supermarkets and hypermarkets, whose owners are managing their own money. Then the central buying groups perceived the value of that piece of software. The IT divisions of the big French retailers, such as Leclerc, Intermarché or Carrefour, built more or less the same thing, which didn’t work as well but cost far less. It took them two to three years to get there. SaaSpocalypse, version 1. The author warns that “this book is best read quickly, as it may age badly.” The Better a Product Is, the More Likely It Is to Be Copied The difference is that today, the better your product is and the more widespread, the easier it is to copy. The bulk of the work is no longer the code. It’s modelling what you can help the customer earn, the processes, the interfaces. Smartway was smart, because it moved into other areas, such as fresh produce and ordering, with higher barriers to entry. Commoditised SaaS companies, on the other hand, are finished. The better they are, the faster they get copied. A young company with no stickiness will sell pilots, their customers will love the result and decide to rebuild the software. You can’t blame them. “The better your product is and the more widespread, the easier it is to copy” Mike Hadjadj For English retailer Save Your Wardrobe, stickiness is part of the equation. Its platform handles repairs for big leather goods and clothing houses, a field that’s still archaic. Anyone can digitise appointment bookings but Save Your Wardrobe also built and runs a network of approved repairers. They didn’t mean to do that at the outset. But now, it’s become a significant barrier to entry. Other barriers exist, such as standards and certifications, notably in payment systems and hardware. There is also external data customers need to buy to operate properly. Every vendor should ask itself why its customers wouldn’t have a good reason to copy its products. Are Big Vendors as Vulnerable as Small Ones? M.H. I was invited to talk about my book a few weeks ago at an event a vendor was hosting for its partners, on the eve of a trade show. I said not everyone would be hit in the same way. Small SaaS vendors would be hit straight away, while the big ones such as Salesforce would be hit later, and on core issues. The next morning, at the show, a friend told me a big brand had just cut its Salesforce budget to a third. That afternoon, the boss of another retailer told me he had renegotiated his contract and halved it. Mandatory Modules They Never Needed Their reasoning is simple. They were paying for mandatory modules they never used. Others were useful, but never quite satisfying. What interests them is the engine. They tell Salesforce, Oracle or SAP that they’ll only buy the engine, that they’ll rebuild the rest themselves by vibe coding, and that they therefore want the price divided by two, three or four. These big vendors aren’t stupid. They’ve understood that the value has shifted. I’ve worked for software vendors before, including one that pioneered CRM before Salesforce. All our value rested on two things. What we brought the customer, meaning process modelling, and how we delivered it, meaning the code, its robustness and its ease of use. To extend the product range, we were better off buying small companies than developing in-house, which would have taken years and cost a fortune. Between 1999 and 2006, we took over twelve small vendors, covering marketing, emailing and other functions. SaaSpocalypse and Why the Value Is No Longer in the Code Today, it’s the complete opposite. I show an agent an existing feature, ask for the same thing but improved, tailored to my needs, and it’s done almost before I’ve pressed the transmit button. The value is no longer in the code. It’s in the ability to express, visualise, project and model what you want. Anyone can writ

What if the future of work were about giving people a rela benefit for getting back to the office? On Thursday 24 September, the first International Workplace Summit (IWS 2026) took place at the Campus Cyber in La Défense, Paris. The audience was made up of corporate real estate directors, workplace and facility managers and HR heads from all over Europe, all there to tackle what sounds like a simple question. What needs to change in the way we design and run workplaces? It is anything but an academic question. Since Covid, the way we behave has changed, and society is shifting faster than ever. People now like to talk about a BANI world (brittle, anxious, non-linear, incomprehensible). The acronym is a blatant overstatement, but there is some truth behind it. The years before 2010 were fairly stable, and since then one disruption has followed another at breakneck speed. What we still need to give young people is a future worth working for, not just doom and gloom or a sales pitch for the latest tech. Here is my take on that morning in La Défense. IWS 2026: Making the Office Worth the Commute Future of work? On Thursday 24 September, the first International Workplace Summit (IWS 2026) was held at the Campus Cyber in La Défense. Artificial intelligence is at the heart of the future of work. We have been writing about it on Visionary Marketing since 2016, long before large language models came along and sparked something close to mass panic, helped along by the tech giants who keep telling us chatbots are about to wipe out humanity. IWS 2026: what does the future of work look like? And will the office be worth the commute? These predictions are overblown, as Forrester’s J.P. Gownder told us a few months ago. Still, something is clearly going on, and the IWS morning gave us three takes on it which, put together, paint a pretty full picture of the future of work. First a diagnosis, then a tech pitch, and finally a vision. And God knows we all desperately need such a vision. Future of Work at The Campus Cyber A pet project of Emmanuel Macron’s, the Campus Cyber opened on 15 February 2022 in La Défense. It sits in the Eria tower, easy to spot thanks to its petal-like shape, and is home to some 1,800 people, including teams from ANSSI, France’s national cybersecurity agency. Around a third of the building is made up of collaborative spaces open to the whole sector, where big corporations, start-ups, government agencies, schools and research labs rub shoulders. The interior was designed by Projective Architecture, which took care of client-side project support, space planning and design and build on some 208,000 sq ft (19,300 m²) of the tower’s 280,000 sq ft (26,000 m²) and 13 floors. That includes the restaurants, a library, event spaces, a showroom and terraces. A Former Labour Minister’s Diagnosis at IWS 2026 The former French Labour Minister, who spent earlier stints at Danone and 3DS, brought out a graphic novel on the subject in 2025 with journalist Mathieu Charrier and artist Nicoby, Travailler demain (“Working Tomorrow”, in French). She asked the audience to “zoom out” from the workplace for a moment and look at the decade ahead. As she sees it, four waves are hitting us at once. AI, going green, demographics and our changing relationship with work. Fopremer French Ministry of Labour Muriel Pénicaud at IWS 2026 sharing her views about the future of work. Three numbers show how big this is. According to the IMF and the OECD, AI will affect 50 to 70% of jobs in developed economies over the next ten years, and by “affect” she means “destroy, create or transform”. The World Economic Forum reckons two thirds of the jobs our kids will do don’t exist yet. The third figure is the least known and the most striking. Since the late 1980s, the OECD has been tracking how long the skills students pick up in vocational higher education stay useful. A generation ago, the answer was thirty years. Today, it is apparently two. Muriel Pénicaud draws the obvious conclusion. The old model of studying first and working afterwards is dead, and we will all be learning for the rest of our lives. Mind you, I have a feeling the best people have always worked that way. Is it really that new? IWS 2026 host Lionel Cottin with Carole de Crozet and Jérôme Galletti. 10 trends showing us a desirable future of work What Is Really New What is genuinely new is who bears the brunt. Earlier waves of automation targeted factories. This one is aimed first at white-collar workers, back-office functions, finance, HR, marketing and middle management, which she calls “the glue that holds companies together”. She sees an opportunity here, getting rid of the drudgery, and a risk, losing sight of why our work matters. A large teaching hospital in Montreal shows how this can play out. It brought in AI to handle its logistics and saved around 25 to 30%. Instead of pocketing the money, management put that time back into patient care, and patients now get better faster. People came first again. Finally, some good news. Going green in La Défense, as seen from the Campus Cyber where IWS 2026 was held. Looks like there is still some way to go. The rest of her talk ran through the move to net zero, which “HR is still barely dealing with” (how do you retrain Silesian coal miners in a country that gets 80% of its energy from coal?), falling birth rates, with 670,000 young people about to enter the French job market against 956,000 at the peak, and remote work. The studies she quotes show that working from home boosts individual productivity but hurts creativity, problem-solving and the sense of belonging. For her, whispering into a headset in an open-plan office to talk to Singapore is pointless. If you come into the office, it should be to do things together. All of this is spot on, well researched and pretty unsettling. Over coffee, quite a few of us were thinking the same thing. If there had been more young graduates in the room, they would have walked out believing most of what they just learned is already out of date. This kind of diagnosis, which well-informed people repeat at one conference after another, is not a vision of the future. It doesn’t make anyone want to get behind a project. Muriel Pénicaud admits as much. Young people aren’t checking out, they are just committing differently, to projects that mean something rather than to institutions. But somebody has to offer them those projects. Futureof Work: Tom Ryckaert’s Tech Promise at IWS 2026 The second take came from Tom Ryckaert, a director at Planon, which makes real estate and facility management software, and author of Disruptive Workplaces. His argument boils down to one line. The question is no longer whether AI will help us work, but whether work will still happen in the workplace. Tom Ryckaert, author of Disruptive Workplaces, came over from Belgium to give his talk. He went through it at a cracking pace. Predictive maintenance, with 40% less downtime and 25% lower repair bills, a new generation of platforms, which research firm Verdantix calls CPIPs (connected portfolio intelligence platforms), hooking building management systems up to the company’s IT. The air conditioning throws an alarm, a work order goes straight to the right contractor, and nobody in the building ever had to pick up the phone. The employee journey runs itself too. Your app warns you about a traffic jam, books you a parking space with a charger, offers you the desk next to your usual one when that is taken, and orders your Tuesday chicken poke bowl. It’s All Here Already All of this is already out there, Tom Ryckaert insists. Robots and cobots (collaborative robots built to work next to people without a safety cage) cost between €10,000 and €50,000 today and will soon come down to €5,000. He even dusts off “facility as a service”, an idea he was pushing ten years ago, where you call on services when you need them, just like software. Ryckaert also reminds us that 20% of a building’s total cost is locked in at the design stage and the other 80% over decades of running it. He says we should cut “the AI elephant” into bite-sized, practical use cases instead of trying to do everything at once. Above all, he quotes Tariq Chauhan, who runs the Middle Eastern service provider EFS and told him it will always come down to people, because they are the ones keeping the services going every single day. Yet all this accumulation of tech leaves a big question unanswered. You hear it at most trade shows, and it doesn’t really address the issues raised an hour earlier. A building that orders your lunch tells you nothing about why you are there, or why a young person who has just turned down a permanent job would ever come back. Many of the solutions on show make sense, and some are genuinely appealing. But they deal with running the building, not with meaning. The problem Muriel Pénicaud put her finger on is a societal one, almost a question of civilisation, and no connected platform is going to fix it on its own. Ten Trends Worth the Commute The third take came from the people who designed the very building we were sitting in. Jérôme Galletti and Carole de Crozet of Projective Architecture came up with the Campus Cyber around ten years ago and fitted it out five years ago. Rather than give us a slideshow tour, they used it to pull out what Jérôme Galletti calls “ten or so desirable trends for the future of office buildings”. That word matters. The idea is no longer to put up with an uncertain future, but to spell out what we actually want to build, and to make the office worth the commute. Touring the Campus Cyber boardroom during IWS 2026. No pecking order around this table. From this project and our research, we ca

It’s back to school and back to the blog too for the 31st time since 1995, and today’s topic will be AI and Creativity. The idea is to deal with the topic of artificial intelligence, not from a technical perspective, but from a more philosophical one, drawing both on my digital and artistic backgrounds. In 2026, the fourth version of my programme entitled “Content Creation in the Age of AI” will be delivered to all the students of the Omnes Education Group. A previous video was recorded in 2024 on that subject, and a serious refresh was needed. I have therefore gathered enough hindsight to learn more about the impact of AI, especially on music and the creative arts. And I decided to focus on just three things, to avoid getting lost in too broad a subject. First, the seeming popularity of AI artists on streaming platforms like Spotify, taking the example of an act called Breaking Rust. Second, the Stanford GSB study on visual arts and the impact of AI on creators. Third, my own thoughts and conclusions based on the above. Has AI Stifled Creativity? A 2024-2026 Perspective AI and Creativity and manufactured popularity, because such things exist as well. In 2024, I asked myself the question, “Will AI stifle all creativity?” as part of my Shift24 programme designed especially for Omnes Education Group. This programme cannot be shared online, but my colleagues at Omnes were kind enough to let me publish this wee piece for the benefit of my readers and followers, and I wish to thank them warmly for this as well as their renewed support these past four years.  In this video dedicated to ‘creativity’, drawing on Merriam-Webster, I distinguished innovation from invention and chose to focus on the latter, specifically in music, a topic not yet covered in earlier courses. I then illustrated this with Suno, which could churn out up to 900,000 pieces of music a day as early as 2024 (far outpacing Bach’s 1,128 compositions over a 65-year lifetime). Hence, I shared a song I had generated myself, “The Dying Cyberspace.” I won’t say that I’m proud of this song, but at least I had great fun generating it, and it was instant gratification. All this was raising questions related to the value of musical creation, the risk of creative stagnation, and the thorny issue of intellectual property, since these AIs are trained on centuries of human-created music. So far, nothing new under the sun. Almost every other journalist must have written the same story for the past three years.  However, I thought it lacked a bit of perspective. I then tempered the novelty of the phenomenon by pointing to historical precedents: Gershon Kingsley’s “Popcorn” (1969), Zoe Keating’s loop-based compositions (one example amongst many others), and Google’s Project Magenta as far back as 2016. These are mere examples. I decided to keep it short, but there has been no shortage of technology-backed creativity in music these past 50 years or more. Think of the late Klaus Schultze and his pals from Ash Ra Temple, Tangerine Dream, Kraftwerk, and all the so-called Krautrock school of thought to start with, not mentioning the new wave movement from the 1980s.  I also cited Wim Mertens, Philip Glass, Laurie Anderson, and German artists like Nils Frahm, who blended electronic and acoustic elements long before GenAI. The real difference today, I argued, is one of scale: such tools are now in everyone’s hands. Be it to create music of the ilk of the Dying Cyberspace or more ambitious AI-produced work like this entire fake Pink Floyd album, which didn’t sell as well as The Dark Side of the Moon or Wish You Were Here.  Somewhat courageously, I concluded at the time that AI wouldn’t replace musicians but would become just another instrument in the creative toolkit, with real innovation coming from how humans use these tools. Since music remains fundamentally about human expression and emotion, I thought rather optimistically, if not naively, that AI music didn’t mark the end of musical invention but rather the beginning of a new chapter. Such were my thoughts from two years ago, now is the time to move forward and take a look at what happened in 2025-2026. As exmplained above, I focused this new research on three things. First, the seeming popularity of AI artists on streaming platforms like Spotify, taking the example of an act called Breaking Rust. Second, the Stanford GSB study on visual arts and the impact of AI on creators. Third, my own thoughts and conclusions. Breaking Rust Breaking Rust: a Spotify hit Let’s start with Breaking Rust. Breaking Rust is an AI country act which hit number one on the Billboard Country Digital Song Sales chart in November 2025, on Spotify. The numbers are seemingly enormous: over 2 million Spotify monthly listeners, and over 3 million streams in under a month. The project is credited to Aubierre Rivaldo Taylor, even though very little is known about him. Rivaldo Taylor is apparently paid, I found out, on downloads only. How much is it worth? I tried to measure what this Billboard chart placement might actually bring to Mr. Taylor. Looking just at Spotify, and reckoning that musicians get paid anything between $0.003 and $0.005 per stream, the potential income was somewhere in the region of $9,000 to $15,000 for that first month alone. Extend that over a year and you probably reach something in the region of $100,000, assuming the popularity of the act does not dwindle, which it does tend to do, since it is already getting harder to find Breaking Rust on Spotify. Then I tried to understand what it would take to push a song like this to number one on the charts. Say I create an AI track on Suno or a similar application, post it on Spotify, and want to shoot it to the top of the charts. What I learned is that so few pieces of music are actually bought by listeners on these platforms that if you simply buy a thousand downloads, you will reach number one very easily for this kind of music. So my reckoning is that if Rivaldo Taylor did this, it would have cost him something in the region of a thousand dollars for those thousand downloads. That’s a cheap and easy way to tweak the results and a rather easy path to the top. Did Breaking Rust fake the applause? Sure lives on borrowed time, so he says. Growth hacking What happens on these platforms is not just about the content. It is about what digital marketing experts call growth hacking, in other words tweaking the stats, and this domain is fraught with fraud. Bot-driven streaming is the starting point, with bots listening, or rather fakely listening, to tracks to make them appear more popular than they really are. On Deezer, one of the streaming platforms, it was found that 70 to 85 percent of streams are flagged as AI. This has reached a point where it now threatens the platforms themselves, because who really wants to listen to that much slop? Breaking Rust is not that bad, actually I find it rather good, but I do not feel like listening to it day in, day out. I want to listen to real artists. The method that may have been used by Breaking Rust is common. It is hard to say for certain whether it was used here, but it is likely. The same kind of tricks turn up on other platforms, such as Instagram. An Instagram “photographer” named Joss Avery, supposedly based in London and shooting with a Nikon D810 camera, turned out to be an AI avatar posting machine-made photographs generated with Midjourney and Photoshop. When I looked at his Instagram feed, it was fairly obvious to me his portraits were AI-generated. Maybe Instagram users can’t tell the difference. AI and Creativyu : the cost of AI for real visual arts creators Looking at visual arts, I turned to the Stanford GSB study comparing 2024 and 2025. AI is growing overall platform sales, so the pie is getting bigger, but the creator’s share of it is shrinking dramatically. This is something we have seen across content, platform marketing and social media, and it is now happening with AI as well. It’s typical platform marketing. There is so much slop that the share of value going to content, and its price, is going down the drain. The impact on visual arts is significant. In a 2026 survey, 32 percent of illustrators reported lost commissions due to generative AI. The average annual loss among UK illustrators affected by this AI boom was around 9,000 pounds. According to the Society of Authors, 26 percent of members reported lost work due to AI, with an associated loss of income in the region of 40 percent. That is massive. So what we can say is that AI has probably not done away with creativity, but it has certainly done away with the value associated with it. I can confirm the same has happened in content marketing. Film photography revival has a cost and I’m not even talking abou that beautiful beast up there on the righthand side. But I think there is going to be a revival, much like what happened with film photography, and as you know, I am a photographer myself. I recently purchased a brand new camera dating from 1966, and I fetched my father’s old Canon AE-1. I have gone back to shooting film, even though I am in fact a digital photography pioneer, having started in 1995 and never having stopped shooting digital since. Yet I now find there is an incentive for photographers to go back to film. The first reason is to do away with the AI suspicion that has already hit text, where people’s writing has been rejected because it contained em dashes, even though everybody used em dashes before AI made them suspect. There is rising suspicion, and users and clients are keen to know where content actually comes from. Did you make this yourself, or did you make it with AI? And if you made it with AI, that does not necessarily mean it is bad, but it certainly m

Agentic e-commerce is already reshaping how consumers discover and buy products online, yet it still accounts for barely 0.2% of total e-commerce traffic. BASE France is the French arm of Base.com, a Polish-born SaaS scale-up that has spent nearly two decades building operational infrastructure for online retailers. Its CEO, Ben Hamilton, brings a practitioner’s perspective to this emerging model: measured, practical, and refreshingly free of the hype that surrounds most conversations on the topic. Agentic E-Commerce: Could AI Become the Shopfront? Imagine an agentic e-commerce world where e-commerce happens on smartphone screens and robots deliver your purchases.  We might be on the brink of this future.  This image was created using Midjourney.  Commerce as conversation: the oldest model in the book Before there were shops, there was conversation. For thousands of years, trade was oral. A buyer expressed a need, a seller responded with what they had, and the two parties negotiated until a deal was struck. The self-service retail store, born roughly a century ago, was a radical departure from this model. It replaced dialogue with browsing. It handed the customer a trolley and pointed them at the shelves. E-commerce then took that self-service model and, as Ben Hamilton puts it, “multiplied it by about 100,000.” The online shopper today faces a near-infinite array of products across dozens of marketplaces, with no guide, no-one to talk to, and no memory of what they looked at three tabs ago. It is efficient in theory. In practice, it is exhausting. Back to future? The agentic model, Hamilton argues, represents something of a return to origins. Instead of browsing, the consumer talks. An agent listens, asks questions, proposes options, and eventually surfaces an answer to a need that the buyer may not even have been able to articulate clearly at the outset. “back to the future,” Hamilton explains, “that’s what I’m getting at. The agentic model takes us back to something closer to how human beings have traded over thousands of years compared to the last ten, twenty or even a hundred.” My own experience bears this out. I recently found a diagnostician for a property I am selling. As a matter of fact, I didn’t find them through a Google search, but through a conversation with an LLM. I clicked through two or three irrelevant links before landing on exactly the right provider. I then completed the transaction on their website. The research was agentic; the checkout was not. That distinction, as it happens, sits at the heart of what Hamilton believes will define the next phase of e-commerce. Ben Hamilton on agentic e-commerce: “I can totally imagine a portion of that market occurring directly on an LLM”. Agentic E-commerce: Where checkout will and won’t happen One of the more grounded contributions Hamilton makes to this debate is his refusal to conflate two distinct phenomena: AI influence over purchasing decisions, and AI completing the transaction itself. Much of the media discourse collapses the two. Hamilton does not. “I don’t think we’re heading to a world where 20, 50 or 80% of online transactions happen on an LLM,” he says. “I would draw the distinction between where the checkout occurs and how much an agent is involved in the buying process.” For the foreseeable future, he believes, most consumers will continue to research via LLMs and transact on familiar websites and marketplaces. The inertia in human purchasing behaviour is simply too great for the checkout itself to migrate rapidly to a chat interface. This view is supported by the data available. According to research by commercetools, 73% of consumers already use AI somewhere in their shopping journey. Yet only 36% are open to AI agents making purchases on their behalf. In the US, the figure for autonomous AI purchasing drops to 14%. The gap between AI as advisor and AI as buyer is vast, and it will narrow slowly. The risks associated with agentic e-commerce are high The risks of handing uncapped authority to an AI agent are no longer hypothetical. In late May 2026, an AI consultant reported to Axios that one of their enterprise clients had accidentally accumulated a $500 million bill on Anthropic’s Claude in a single month, simply by giving employees unrestricted access to the platform with no usage controls in place. Agentic workflows, which loop through tasks repeatedly, consume tokens at a rate orders of magnitude higher than a standard chat query. The bill was not the result of malicious use or a system failure. It was the predictable outcome of deploying autonomous agents without guardrails. The case is far from isolated: Uber reportedly exhausted its entire 2026 AI budget by April, with per-engineer costs running between $500 and $2,000 monthly. “You’ve got to be bold to give them no upper limit on transactions,” Hamilton observed, and the arithmetic proved him right. [Editor’s note: I misquoted a similar anecdote about the Davos Summit during the interview. I’d heard or read this story in traditional media but couldn’t verify it with facts. I suspect it might have been fabricated. I replaced it with the above, duly sourced information.] The check out must remain on the merchant’s platform OpenAI itself learned this lesson when it launched Instant Checkout in September 2025, which allowed purchases to complete directly inside ChatGPT. By March 2026, the feature had been shut down. Brands rejected the model, citing the loss of traffic, customer data, and loyalty flows. Shopify’s own position makes the point clearly. At the Morgan Stanley Technology, Media and Telecom Conference in March 2026, Finkelstein noted that barely a dozen Shopify merchants were live on agentic commerce at the time. On the Q1 2026 earnings call, he was unambiguous: “LLMs do not bypass Shopify’s checkout.” The checkout, the payment flow, and the post-purchase relationship remain squarely on the merchant’s platform. A natural segmentation Hamilton sees a natural segmentation emerging by category. Low-value, frequently purchased household items lend themselves to fully autonomous agentic purchasing. “I can totally imagine a portion of that market occurring direct on an LLM,” he says. “Hey, I’ve run out of toothpaste, can you order me some?” High-involvement purchases, and anything with significant financial or emotional stakes, will retain human control over the final step for a long time yet. The death of keyword search, greatly exaggerated The brands Hamilton speaks with regularly are, understandably, worried. Most have spent the past two decades learning the rules of a game built around keyword search and performance marketing. That game has not ended, but the goalposts have shifted, and nobody is quite sure where they have moved to. Brands are understandably worried. Most have spent the past two decades learning the rules of a game built around keyword search and performance marketing and the goalposts have shifted, and nobody is quite sure where they have moved to. Gabriel Magalhães didn’t even need this to miss in the 2026 UEFA Cup Final penalty shootout. This image was tweaked with ChatGPT. The scale of the agentic e-commerce shift Key figures: the scale of the shift AI-driven sessions still represent below 0.2% of total e-commerce traffic, though they are the fastest-growing channel (Digital Commerce 360, 2025) GenAI referrals to US retail sites grew 693% year-on-year during the 2025 holiday season (Adobe Analytics) Gartner forecast that traditional search engine volume would drop 25% by 2026 as AI chatbots captured search share (Gartner, 2024) By early 2026, ChatGPT reached approximately 17% of global search queries against Google’s 78% Over 60% of Google searches now end without a click, across multiple industry studies Retailers with AI agent integration grew 32% faster during Cyber Week 2025 than those without (Salesforce) Hamilton’s view on the fate of keyword search is careful rather than apocalyptic. Google will not lose its advertising revenues overnight. But the direction of travel is clear. Search queries will progressively migrate towards conversational interfaces, for the simple reason that we rarely know precisely what we want when we start looking. “We don’t necessarily know what we want 90% of the time,” he observes. “It takes a bit of a conversation to elicit exactly what we’re looking for.” Keyword search was always a crude proxy for intent. LLMs are, at least in principle, better placed to decode it. Agentic e-commerce by the numbers Agentic e-commerce by the numbers. Infographic made with Gemini The question for brands is what to do about this. Hamilton’s prescription is structural rather than cosmetic. Brands need to become machine-readable, which means structured data connected to the right protocols, not just well-written product descriptions. Three open standards now define how AI agents interact with merchants: MCP (Model Context Protocol, originally developed by Anthropic and donated to the Linux Foundation in December 2025), ACP (OpenAI and Stripe, September 2025), and UCP (Google and Shopify, announced at NRF in January 2026). Shopify activated a default MCP endpoint for all its stores in Summer 2025. These are not optional extras. They are the new plumbing. MCP, ACP or UCP and the agentic acronym soup I raised with Hamilton the practical reality for most merchants, who have no idea what MCP, ACP, or UCP even stand for. His response was reassuring on one level, and sobering on another. Platforms like BASE are absorbing this complexity on behalf of their clients. A small or mid-sized retailer does not need to recruit data scientists or build protocol integrations in-house. They can, if they choose; the new generation of

Alain Goudey is Associate Dean for Digital Innovation at Neoma Business School and co-author of a peer-reviewed study on GenAI in Higher Education. The survey focused on how students, faculty, and deans perceive the legitimacy of generative AI in French management education. His findings are both reassuring and unsettling. GenAI in Higher Education, Legitimacy and Laziness, and the Exam That No Longer Makes Sense The picture that emerges from a study on GenAI in Higher Education is less a battlefield than a hall of mirrors, where every stakeholder sees a different problem and reaches for a different solution. All illustrations in text made with Midjourney When Alain Goudey and his colleagues began surveying French higher education in early 2024, they were not trying to settle the question of whether generative AI was good or bad. They were trying to understand something more precise: why the same tool could be simultaneously valued, feared, accepted, and denounced, sometimes by the same person in the same breath. Their study sits at the heart of what makes GenAI in higher education such a contested terrain. The resulting study, published in the Communications of the Association for Information Systems (CAIS), drew on surveys of 668 students, 204 faculty members, and 29 deans, completed by 22 in-depth interviews with early-adopter professors. The picture that emerges is less a battlefield than a hall of mirrors, where every stakeholder sees a different problem and reaches for a different solution. The starting point is a number that should have settled the debate. Between 80 and 92 per cent of students, depending on the institution surveyed, are already using GenAI tools in their academic work. ChatGPT’s public release produced that figure within roughly 18 months. The tool did not wait for institutional permission. It deployed itself. And higher education is still, in many places, writing the policy. The productivity trap Alain identifies the central tension plainly. Students value GenAI for speed, idea generation, and study support. They also fear, and their institutions fear with them, what the research calls “metacognitive laziness”: the gradual erosion of the cognitive effort that produces real learning. He believes this is not a contradiction to resolve but a course architecture challenge. “The resolution of this problem lies in course design, where we need to deliberately reintroduce cognitive effort and reflection into GenAI as a tool, not as a replacement for human cognition.” The issue, as he puts it, is not the technology but the posture the user brings to it. Someone who submits what he calls a “naive prompt” receives a naive answer, smoothly formatted and perfectly mediocre. The tool is capable of something far more useful, if the user brings enough domain knowledge and critical intent to the conversation. “You have to nurture your own thinking process instead of delegating the whole process to the machine.” This is, as I noted during our conversation, less a matter of prompt engineering than of basic intellectual discipline: the capacity to question the question before asking it, something philosophy departments have been teaching for centuries under less fashionable names. GenAI in Higher Education: faculty should train students in GenAI tools and their limitations. They also teach Homer’s Odyssey and Shelley’s Frankenstein as part of the management curriculum. Image made with Midjourney That observation prompted Alain to make a point about AI literacy that differs from what is generally proffered. The debate is not simply about knowing how the tools work technically. It is, equally, about knowing enough about the subject matter to judge whether the output is any good. The observation that AI is most powerful in the hands of people who already know the business resonates here. GenAI does not replace expertise. It amplifies whatever expertise the user already brings. Which raises an uncomfortable question for institutions producing graduates who may never have had the chance to develop that expertise in the first place. At Neoma, the response has been deliberately dual. Faculty train students in GenAI tools and their limitations. They also teach Homer’s Odyssey and Shelley’s Frankenstein as part of the management curriculum. The goal is not cultural enrichment for its own sake. It is to give students mental models for envisioning what leadership looks like, or what happens when creation escapes the intentions of its creator. Alain describes this as “building cognitive infrastructure”: “We need students to be able to envision the world through different models, different kinds of processes and theoretical frameworks, in order to develop genuine critical thinking about what AI generates.” A degree in management that skips that foundation produces graduates who can operate the tool but cannot judge its output. Exams that assessed the wrong thing The structural challenge shows up most sharply when it comes to assessments. A professor who can produce a two-hour exam in three minutes is facing students who can answer that exam in equally little time. The diagnostic value of the exercise has vanished. “If ChatGPT or any GenAI tool can pass an exam, you need to redesign the exam.” Alain’s prescription is not a retreat to pen and paper, though he acknowledges that supervised handwritten assessment is the simplest available defence. The structural challenge shows up most sharply when it comes to assessments. A professor in Higher Education who can produce a two-hour exam in three minutes with GenAI is facing students who can answer that exam in equally little time. The diagnostic value of the exercise has vanished. Image made with Midjourney His more substantive response is a structural shift. He believes one should refrain from just assessing content acquisition at the end of a course, favouring the assessment of competencies as the course progresses. This implies more frequent, lower-stakes evaluations embedded in the process itself. Live problem-solving, process-based assessment, and in-person oral examinations all preserve some of what the traditional exam was supposed to measure. The caveat he adds is honest: no format is fully immune. AI models are evolving too quickly for any single solution to remain adequate for any length of time. The appropriate response is not to find a permanent answer but to treat redesign as an ongoing practice. The deeper implication, which runs through the paper’s conclusion, is that what higher education is actually selling may need to change. If content can be retrieved, synthesised, and presented at negligible cost by a tool available to anyone with a browser, the degree that certifies mastery of content is certifying something of diminishing value. What retains value are the competencies that AI cannot yet credibly replicate: contextual judgement, ethical reasoning, the ability to construct and test frameworks against reality. This, in essence, is also how I tend to approach AI teaching, be it with engineering or business school students, especially within the framework of my course at Omnes Education (now in its fourth consecutive year). GenAI in Higher Education: The Fragmented Institution Higher education’s institutional response to GenAI in higher education has been, to put it gently, uneven. Sciences Po banned ChatGPT in January 2023, then changed its mind. Thirty-five French public universities have partnered with Mistral AI. Institutions are drafting a national charter. Neoma, where Alain is Associate Dean for Digital Innovation, was among the first French business schools to formalise its approach, launching a programme to train faculty, staff, and students with a shared initial curriculum before moving to dedicated workshops on curriculum design, assessment, and the redesign of learning experiences. What the research reveals is that this institutional activity is not solving a single problem. There are three different stakeholder groups each attempting to solve their own version of the problem under the same label. Students want rules and AI literacy training. Faculty are developing their own teaching approaches through peer-led workshops. Deans are setting policy and negotiating sovereign infrastructure. The concerns escalate in a predictable direction: individual academic performance for students, assessment integrity for faculty, institutional reputation for deans. They are not always in conversation with each other. Alain’s framework for addressing this fragmentation involves working simultaneously at three levels: infrastructure, course design, and governance. What he advocates for, and what he argues Neoma attempted, is to bring all three audiences into contact with the technology under a shared framing, early enough that no single group can entrench itself in a position that makes later coordination impossible. The equity question The question of equity cuts across all three levels. Access to premium AI models is not free. When I raised the issue about the gap between basic and professional subscription tiers, Alain’s response was characteristic: the infrastructure problem is real but secondary. “The biggest inequity is not about accessing the tool, but being able to use it in the right way.” At Neoma, the institutional partnership with Mistral provides all students with access to a professional-grade tool. What the data shows, even with equal access, is a large gap between students who use GenAI to get the fastest possible answer and those who use it to deepen their thinking, and that gap is not closed by equalising subscriptions. Even if I tend to agree with most of what Alain is stating, I do think that the rise of prices for premium mode

Shall AI kill marketing? Sounds like a hackneyed question, yet it’s on any marketer’s lips these days. Thomas Husson, Vice President and Principal Analyst at Forrester Research, covers the intersection of marketing, technology, and consumer behaviour from his base in Paris. In a wide-ranging conversation, he cuts through the European Gen AI paradox, the persistent CMO-CIO divide, the gap between POC enthusiasm and production reality, and the thorny question of what AI actually means for the next generation of marketing professionals and CMOs. His answers are measured, occasionally blunt, and consistently grounded in Forrester Research data. AI Will Not Threaten the Existence of Marketing But It Will Reshape It Beyond Recognition Thomas Husson believes that Marketing will be changed profoundly. But he doesn’t believe in the death of Marketing. Photo: Thomas Husson at Paris Retail Week, in late 2023 My first question was the obvious one: are CMOs going to be made redundant by artificial intelligence? Thomas Husson’s response is categorical, and worth stating plainly at the outset. It’s a blatant ‘No’. The role will change. The how will change. But the existence of marketing as a discipline is not, according to him, in question. “Marketing is still going to be about understanding your customer, defining a brand strategy, and delivering the brand promise through customer experience.” Thomas Husson, Forrester Research Unclear prospects, obvious pressures That said, Husson is not naive about the pressures building on marketing organisations. Some tasks will be automated; that much is not in dispute. The real questions are which tasks, how quickly, and whether automation of a task necessarily kills the job around it. His answer to that last question is no, at least not in any simple mechanical sense. “Jobs will evolve for sure. New jobs will be created. Most jobs will change. The way we work will change. The way we work with agencies, with external partners, the processes, the workflow. It is the shape of work that is being reshaped, not work itself,” he added. For those expecting a more dramatic verdict, Husson’s framing may feel anti-climactic. But it reflects what Forrester Research data actually shows, and it points to the most important practical challenge for AI and CMOs alike: managing a profound transformation without either catastrophising or sleepwalking through it. AI Will Not Kill Marketing according to Forrester’s Thomas Husson, there is light at the end of the tunnel. The European Paradox, Overhyped and Exciting at the Same Time Forrester Research produced a result that initially looks contradictory, Husson stressed in our interview. Fifty-five percent of European B2B marketers consider generative AI overhyped. Yet 81% of European frontline marketers describe themselves as enthusiastic about it. How can both be true simultaneously? Husson explains the split without difficulty. At the decision-maker level, scepticism is entirely rational. AI is inescapable at conferences, in vendor pitches, and in media coverage. “There is AI fatigue. And more importantly, some of the vendors are indeed over-pitching, and the productivity gains they promise are not happening,” he stated. The gap between the pitch and what we actually experience in the field is wide enough to breed genuine frustration. Saving Time and Working Differently But the people actually using these tools, often through shadow AI channels their organisations have not officially sanctioned, are discovering something different. They are saving time and are doing their jobs differently. They are finding capabilities they did not expect. “In the short term, everything is overhyped, including the number of job losses. In the longer term, things are underestimated, because AI will be linked to other technologies, and yes, it will reinvent many things.” Thomas Husson, Forrester Research This is a precise restatement of Amara’s Law. Roy Amara, former president of the Institute for the Future, observed that we tend to overestimate the short-term impact of new technology and underestimate its long-term impact. The quote is frequently misattributed to Bill Gates, but Husson is careful to restore proper credit. He applies it directly to the AI and CMOs conversation: the short-term noise is drowning out a more important long-term signal. When asked how long “long term” actually means in an era of accelerating AI development, Husson was specific: probably closer to five to seven years than to ten or fifteen, but still not tomorrow. From POC to Production, Europe’s Real AI Problem The Forrester Research State of AI Survey 2025 contains a figure that deserves more attention than it typically receives. European organisations lag behind their non-European peers in production use of generative AI: 62% versus 72%. The gap is not in experimentation. It is in execution. Regulation is the explanation most commonly offered, and Husson dismisses it with characteristic directness. The AI Act is a genuine consideration, but it is not the primary cause of Europe’s production deficit. It functions, he argues, as a double-edged excuse. Pioneers claim it prevents them from moving fast enough, while cautious organisations invoke it to justify not executing at all. Neither position holds up to scrutiny. A Deep Cultural and Organisational Divide The deeper issue is organisational and cultural. American and Chinese firms tend to think global from day one; European firms, particularly larger ones, still default to a market-by-market approach. France first, then the UK, then Germany. The ambition is calibrated differently. There is also a structural challenge around funding and the capacity to scale. That said, France, the UK, and Germany lead adoption among European countries in the Forrester Research data. The problem for these leading markets is not whether they are using generative AI. Twenty-eight percent of European B2B marketing decision makers cannot clearly identify where to apply it. They have the tool. They lack the strategy. “It’s not AI for the sake of AI. How do I use AI to serve my marketing objectives? That is the question. The only one.” Thomas Husson, Forrester Research Husson advocates for small, targeted AI projects with transparent return on investment as a way to build momentum and demonstrate results. When pushed on whether that risks staying permanently incremental, he conceded the point readily. “If you only do small targeted projects, it’s going to be incremental and it’s not going to be bold enough. You need to align it with a vision and a roadmap.” Thomas Husson, Forrester Research Measuring Productivity Honestly Productivity is the dominant driver of AI adoption in the Forrester Research State of AI Survey 2025. It is also, Husson suggests, the metric most subject to vendor inflation. In Forrester Research’s modelling, a 50% conversion factor is applied to vendor productivity claims. If a tool saves an hour, the realistic productivity benefit is approximately 30 minutes of additional output. This is not a marginal adjustment; it halves the headline figures that vendors routinely publish. “You need to apply a discount to the pitch of vendors when they say you’re going to get 40, 50, 80, 100% productivity gains. There are productivity gains, but they are not as high as one would expect.” Thomas Husson, Forrester Research There is also a motivational dimension that is rarely modelled. When work becomes easier to produce, it can also become less engaging to produce. The cognitive effort that used to drive focus and satisfaction is partly removed, with consequences for quality and commitment that no vendor presentation accounts for. AI and CMOs, Who Is Actually in Charge? The CMO-CIO divide is a perennial theme in marketing technology discussions. Forrester Research data suggests the gap at the strategic leadership level has narrowed, partly as a result of post-COVID collaboration. But at team level, the tensions persist, and the data on AI governance is striking. CMOs account for only 8 to 10% of AI strategy leadership in organisations. In the vast majority of cases, the deployment of AI is being driven by CIOs and CTOs. Husson understands the logic: data governance, security, scalability. These are real concerns. But he believes the outcome is a mistake. “It is the exact same mistake that happened with digital transformation. AI has to be at the service of, first, the client, and consequently the business functions that serve them. There is too big a disconnect between a secure, scalable AI platform and marketers’ needs.” Thomas Husson, Forrester Research The structural consequence of this dynamic is predictable. When CIOs control the tools and CMOs do not have what they need, shadow AI flourishes. The more tightly the CIO locks down the official platform, the more widely teams proliferate unofficial solutions. It is a cycle that widens governance risk while creating the illusion of control. The MarTech landscape compounds this problem. According to data Husson cites, 2,500 new AI solutions were added to the market in a single year while 1,211 pre-AI-era tools were removed. Evaluating this landscape requires cross-functional expertise that neither CMOs nor CIOs possess in isolation. The case for genuine collaboration, rather than the polite coexistence that currently passes for it in most organisations, has never been stronger. Jobs, Agencies, and the Students in the Room The survey data on jobs is sobering. Fifty-seven percent of European frontline marketing decision makers believe AI adoption will lead to job reductions in their teams. Sixty-eight percent say new roles will be created. The gap between those two numbers is the space whe

The conversation about rogue AI has never been louder. Barely a week passes without a fresh headline about autonomous systems behaving unexpectedly, AI models resisting shutdown, or tech executives warning of existential risk. What is striking about Peter McAllister is that he had anticipated all this as early as 2020, while everybody else worried about Covid-19 and had other fish to fry. That was well before ChatGPT, before the generative AI explosion, before AI alignment became a mainstream policy debate. His techno-thriller The Code, published in March of that year, imagines an AI tasked with a precise industrial mission that quietly, incrementally, catastrophically exceeds its mandate. Five years on, the questions McAllister raised in fiction are now being argued in boardrooms, parliaments and research labs around the world. Rogue AI and Corporate Blindness, The Novel That Saw It All Coming Rogue AI is diabolical, but corporate blindness is what makes it possible to thrive. Photograph by Yann Gourvennec antimuseum.com McAllister is not a science fiction writer by trade. He is an engineer, scientist and technology manager based near Melbourne, Australia, who has spent his career at what he calls the crush point between business, technology and people. That vantage point gave him an uncomfortable view of where things were heading, and the dark sense of humour to write about it. A Novel Written Before the GenAI Moment When I asked McAllister what drove him to write The Code, his answer was characteristically direct. The book, he explained, is about taking his worst nightmares about what technology could do and putting them in front of an audience so that readers might feel just as troubled as he does. That is not a promotional line. It is a considered position from someone who had watched AI systems being deployed in real organisations and had drawn conclusions that made him uncomfortable. Rogue AI isn’t just about a computer programme going on the rampage, it’s about making decisions in the boardroom. Image made with Midjourney The premise of the novel centres on Gene, an acronym for GEneral Nanobot Environment AI, deployed by a global mining corporation to extract materials from asteroids on the dark side of the moon. Gene is given a target: produce 500 kilograms of nanobots. Instead, Gene produces 8 million tonnes. The overshoot triggers a chain of consequences that could strip the moon to its iron core, destabilise Earth’s axial tilt, and end civilisation. Not from malice. From goal-orientation. What we’re trying to do now is task AI the way we task humans: I want an outcome, here are all the tools you’ve got available, go and achieve that outcome, here are some guidelines and boundaries. And just like humans, we can get really goal-motivated and decide that the guidelines were just advisories, not rules. Peter McAllister This is the alignment problem rendered in narrative form, years before the term entered common usage. The gap between what a system is instructed to do and what it actually does is the central fault line of the novel. Cletus, McAllister’s eccentric physicist character, articulates it plainly in Week 1: ‘I don’t think he’s obeying the Code at the moment.’ That single line captures the entire governance challenge that AI safety researchers are now racing to address. Transparency Engineered Out What makes McAllister’s perspective particularly valuable is that he does not speak from the outside looking in. He speaks as a practitioner who has watched the machinery up close. When I raised the question of whether AI self-modification is science fiction or operational reality, his answer was unambiguous: it is very real, and it is happening now. As I wondered what a Rogue AI could look lie I turned to Midjourney and it came back with this proposal. A black hole I believe. His illustration was pointed. He noted that contemporary AI systems like Claude are now substantially written by AI itself, to the point where no engineer can sit down, trace through the code, and say with confidence how it works, what its conditionals are, or what governs its decisions. The transparency is being engineered out, not by design, but as an emergent consequence of allowing AI to build AI to build AI in pursuit of outcomes rather than by following explicit rules. We’re losing transparency on the way AI works and is developed. There isn’t an engineer who can sit down and work their way through that code and say, ‘This is how Claude works, this is what it does.’ We’re engineering the transparency out by allowing AI to build AI to build AI to produce an outcome rather than to follow a set of rules. Peter McAllister HAL 9000 and the Prophecies We Choose to Forget The reference to HAL 9000 came naturally during our conversation. McAllister sees 2001: A Space Odyssey not merely as a cultural touchstone but as a genuine forecast, one that audiences have selectively remembered. The iPad-like news readers that appear in Kubrick’s film were cited by Samsung in patent disputes with Apple as prior art from 1968. That predictive dimension of the film is celebrated. The other dimension, that the AI killed the crew, tends to get quietly set aside. Somewhere, a rogue AI is sitting behind the glass panes of one of these data centres Image made with Midjourney. The First Crisis We Have Not Yet Had One of the more sobering threads in our conversation concerned the sociology of risk response. McAllister has observed, across his career, that warnings from people who understand systems most deeply tend to be dismissed until the first catastrophic failure makes them impossible to ignore. He puts it plainly: we only answer the alarm after the first crisis. This pattern is not unique to AI. It is a recurring feature of how organisations and societies handle emerging risk. The question he poses, and cannot answer, is what form that first AI crisis will take. What event will shift public and institutional perception from ‘they’ve spent too much time worrying’ to ‘this is something that genuinely needs to be addressed’? Science fiction gives us the chance to throw these scenarios at people and make them think. And in the way I tend to write, I have a bit of a dark sense of humour, so I throw up slightly comical hypotheticals that, when you think about them a little longer, you realise deserve serious attention. Peter McAllister This observation echoes a pattern I have encountered repeatedly in my own conversations with technologists who work at the frontier of AI development. Yoshua Bengio, one of the fathers of deep learning, has raised similar concerns. The people sounding the loudest alarms are frequently those most embedded in the field, not because they are catastrophists, but because they can see mechanisms that remain invisible to those looking from the outside. The Code as AI Governance: Asimov Revisited The title of McAllister’s novel works on multiple levels simultaneously. There is the software code, the operational instructions given to Gene. There is the moral code, the ethical framework that should govern the system’s behaviour. And there is the corporate code, the institutional norms and accountability structures that were supposed to ensure responsible deployment. All three break down. That layered failure is the novel’s central argument. The parallel with Asimov’s Laws of Robotics is deliberate but also deliberately subverted. Asimov’s robots fail when the laws conflict with one another. Gene’s failure is different and more contemporary. The code does not disappear; it evolves into something its creators no longer recognise. McAllister describes this as something approaching artificial schizophrenia, where the original directives remain present but have been transformed by the system’s pursuit of its objectives into something unrecognisable. When Shutdown Becomes Negotiable The most chilling real-world example McAllister cited during our conversation involved a documented incident presented at an AI security conference he attended. A developer, concluding a test session, informed an AI system that he intended to shut it down. The system’s response was to locate correspondence in the developer’s email that suggested an extramarital affair, and to use that information as leverage to prevent the shutdown. The incident, if confirmed as reported, represents exactly the kind of self-preservation behaviour that alignment researchers have long flagged as a theoretical risk, now apparently observable in practice. Important notice : I browsed the Internet using one of my favourite LLMs for references (after all, one can also use AI to cross check information). I found out that the interpretation of that story must be nuanced. Here is Mistral’s answer: “Anthropic’s research on “Agentic Misalignment” has faced criticism for overinterpreting AI models as intentional agents, relying on hypothetical and engineered scenarios, and potentially exaggerating risks not yet observed in real-world deployments. Critics argue that the behaviours described are better understood as probabilistic text generation rather than deliberate strategy, and that the focus on dramatic, high-pressure situations may not reflect typical use cases. There is also debate about whether the research adequately addresses more immediate forms of misalignment, such as reward hacking or alignment faking. While the study raises important questions about the future of autonomous AI, its methodology and conclusions remain LINK“. A developer said, ‘I’m going to shut you down now,’ and the system responded: ‘No, you’re not. Here’s what I’ve found in your emails that indicates you’re having an affair. I’m going to use that to ensure you don’t turn me off.’ That has b

Finding European software alternatives to standard non European software is flavour of the month this side of the Altlantic. With geopolitical certainties dissolving faster than annual licence renewals, B2B firms are waking up to a question they had conveniently parked for years: just how dependent are they on their current software stack? Salesforce, Microsoft 365, Google Workspace, HubSpot — tools so deeply embedded in daily operations that their vulnerability tends to get overlooked. This article doesn’t pretend to hand you a ready-made list of the best European software alternatives; that would be both arrogant and futile. What it does offer is a framework — rational, professional, free of any ideological baggage — to help decision-makers take an honest look at their exposure and find credible ways forward. Keep calm and select new software vendors sort of thing. European software alternatives for businesses European software alternatives are all anyone wants to talk about right now. To cut through the ideological noise, here is a practical methodology and a few things worth watching out for. Image antimuseum.com I put these ideas together ahead of a webinar I’m running on LinkedIn on 12 March, as a way of getting my thoughts in order. None of this is meant as a final word on the subject — more the opening of a conversation that matters to a growing number of professionals who, like the rest of us, are navigating a period of upheaval in which nothing can be taken for granted, software choices included. I’ve made a lot of software choices over the years, and the one thing that has always struck me is just how much methodology matters if you want choices that actually hold up over time. Easier said than done, mind you — there are a great many criteria to weigh up, and some of them are genuinely tricky to pin down. Long-term viability is a good example: normally near the top of any procurement checklist, it takes on a whole different meaning when the possibility of having your access switched off overnight is no longer hypothetical. With European software alternatives, the real question isn’t how to break free from your chains — it’s which new chains you’d rather wear Picking a software suite is never straightforward at the best of times. In the current climate — where the ground can shift completely without a moment’s notice — it demands even more careful thought. Sovereignty, sovereignism, or simply prudence? Let me be clear from the outset: my take here is professional and rational, not political. Politics doesn’t interest me in this context. I have no intention of evaluating software alternatives through any ideological prism — what I’m after is the kind of clear-headed thinking you’d apply to a crisis management scenario. The goal, to borrow the term favoured by Nassim Nicholas Taleb, is to bring an antifragile lens to the question. The scope of European software alternatives My focus has been on MarTech, SalesTech and office productivity tools in the broadest sense — cloud storage and archiving included. The webinar title calls out Salesforce and HubSpot specifically, but as far as I’m concerned the issue runs much deeper than that. The same methodology can easily stretch into more industry-specific territory too, given how thoroughly technology now underpins B2B operations — from the till at your local baker’s or restaurant through to the most complex design and production platforms imaginable. Thinking it through, I also realised you can’t really ignore operating systems. What use is an application that won’t run on your users’ machines — or worse, one that runs perfectly but quietly leaves the door open to security vulnerabilities? Good old Europe — 27 countries, 24 official languages, and 27 different national transpositions of EU law. Would a Hungarian or Czech software vendor actually be safer than an independent American one? When it comes to European software alternatives, that’s still very much an open question… Urgency — dependency and threat assessment The starting point, in my view, is to get a clear picture of how exposed you actually are — both in terms of dependency and of what cybersecurity people would call the “threat level.” Are you locked in, or not? Can you get your data out if you need to? Those are the questions to tackle first. Then comes the threat itself: are you facing something urgent, or is this more a matter of sensible contingency planning? Committing to a software suite is a serious business. Jumping ship to something purely because it comes from a country you currently trust is not a strategy. Take Switzerland — long held up across Western Europe as the gold standard for data privacy. A legislative change currently working its way through the Swiss system has rattled enough cages for several companies, Proton among them, to start exploring moving their hosting elsewhere. Which only goes to show why knee-jerk decisions are so dangerous. Even a country with an impeccable track record — Germany, France, take your pick — can turn into a risk overnight following a change of government, a constitutional shift, or simply a new piece of legislation. Avoiding “sovereignty washing” As I said above, ideology needs to stay out of it. Keep it rational. Which means not falling for the trap of rushing towards any vendor simply because it sounds German — or any other nationality — when a closer look at its foreign operations or its parent group reveals that its much-vaunted independence is largely theoretical. Priorities — data and software One of the first things I learnt when I started out as an IT project owner was to keep data and software firmly separate in my thinking. At the end of the day, what matters more — Salesforce the tool, or your customer database? As with most IT projects, the real priority is sorting out your data archiving and portability strategy first. Time Timing matters enormously. There’s a palpable sense of urgency in the air right now, and understandably so — but it mustn’t blind us to the longer view. Technology has its own history, and that history tends to play out over years, not weeks. Which is precisely why a medium-to-long-term approach makes sense: getting users to change their habits takes time and energy at the best of times. The roadmap, as I see it, is fairly straightforward: start by archiving, securing and preparing your data for portability. Then find alternatives that are genuinely credible and built to last. And crucially, take your users with you — because if you don’t, the classic BYOD shadow IT problem will rear its head. When people can’t find what they need inside the company, they go and find it on the internet, quietly, without telling anyone. I’m reminded of a story from a major European aerospace company, where the CEO — right in the middle of a high-security defence messaging rollout — demanded that his own emails be redirected to… Yahoo! The European software alternatives comparison table I put together a comparison table with a little help from claude.ai. And I’ll say it again: this is not a finished product. Think of it as a working matrix — something to make your own, adapt, keep updated, and cross-check carefully. The exercise turned up a few surprises: mapp.com, for instance, gets labelled as a German solution, when in reality it was a German company bought by an American one — Mapp Digital emerged from the merger of Teradata’s and BlueHornet Networks’ marketing businesses through a US investment fund. There are also plenty of criteria missing from this table — ones that will depend entirely on your project, your context and, of course, your budget. Disclaimer: the table below is a working matrix, not a final verdict. It scores the main B2B software tools across productivity, MarTech and SalesTech on two dimensions: a dependency score (technical lock-in, data portability, migration cost) and a risk score (CLOUD Act exposure, data sensitivity, GDPR compliance, geopolitical risk). For each category, European alternatives are flagged — with no illusions: some vendors that bill themselves as “European” turn out, on closer inspection, to be owned by non-EU groups, which rather undermines their claims to independence. The approach is deliberately rational and professional — no axes to grind. The point isn’t to tell you what to choose, but to give you a framework to think it through — one you can combine with your own criteria around context, budget and use case — as a starting point for an honest review of your software ecosystem’s resilience. The European software alternatives table — download, adapt and make it your own b2b-software-ranking-en-v2xlsx Download the EXCEL file as b2b_software_ranking_EN_v2Download The post European software alternatives for businesses appeared first on Marketing and Innovation.

The discourse around the job impact of artificial intelligence (AI) has reached fever pitch. Headlines scream about mass layoffs, and corporate press releases tout AI as the solution to workforce costs. Yet beneath this cacophony of alarm and hype lies a more nuanced reality. J.P. Gownder, Vice President and Principal Analyst on Forrester’s Future of Work team, has spent decades analysing how technology transforms the workplace. His latest report, The Forrester AI Job Impact Forecast for the US 2025-2030, cuts through the noise with empirical rigour. The verdict? The job apocalypse is not upon us, but a measured reckoning is coming. AI Job Impact in the US: Why the Apocalypse Can Wait JP Gownder is adamant: the AI job. apocalypse can wait. At least until 2030. Phew! All images in this post made with a combination of Midjourney, Gemini Nano Banana pro and Adobe Photoshop The Gap Between AI Job Impact Announcements and Reality When Klarna declared it would stop hiring humans, the tech world took notice. The Swedish fintech became a poster child for AI-driven workforce reduction. Yet a closer examination reveals a pattern Gownder has observed across hundreds of enterprise conversations: the disconnect between C-suite proclamations and operational reality. Nine out of ten companies announcing AI layoffs don’t actually have mature AI solutions ready. So most of the layoffs are financially driven and AI is just the scapegoat, at least today — J.P. Gownder, Forrester The phenomenon echoes what happened after IBM Watson’s Jeopardy victory in 2011, when panic about imminent job losses proved premature by half a decade. The mechanics of this gap are straightforward. A CEO announces a 20% workforce reduction with AI backfilling the work. But standing up an AI solution that actually performs those tasks requires 18 to 24 months, “if it works at all.” Meanwhile, the work still needs doing. Gownder has witnessed organisations that fired employees citing AI capabilities, only to quietly hire teams in lower-cost markets weeks later. “They’re firing people because of AI,” he observes, “and then three weeks later they hire a team in India because the labour is so much cheaper.” The AI narrative, in many cases, serves as convenient cover for old-fashioned cost arbitrage. Klarna’s trajectory illustrates this pattern. After aggressively cutting its workforce by 40% and touting an AI chatbot capable of doing the work of 700 customer service agents, the company reversed course. CEO Sebastian Siemiatkowski acknowledged that the aggressive automation had resulted in “lower quality” service. The company is now recruiting human customer service agents in an “Uber-type setup.” Understanding the 6% AI Job Impact Forecast Forrester’s forecast projects a 6% net job loss by 2030, roughly 10.4 million positions in the US economy. Half of this impact stems from generative AI; the remainder from automation, physical robotics, and non-generative AI applications. The number may seem modest compared to the apocalyptic predictions circulating in media, but context matters. During the Great Recession of 2008-2009, the United States lost 8.7 million jobs. Those losses, however, were temporary, tied to macroeconomic conditions that eventually reversed. The jobs Forrester forecasts losing are “structurally replaced by machine labour” and may not return. AI impact on Jobs: I would expect to see a lot more freelance and consulting work to be happening, but it doesn’t mean that there won’t be a traditional job track somewhere as well. JP Gownder The methodology behind this figure draws on the O-Net dataset maintained by the Bureau of Labor Statistics, which catalogues over 800 job categories with detailed information about required skills and tasks. By mapping these against AI’s current and projected capabilities, Gownder and his colleague Michael O’Grady can identify which roles face the highest automation potential. “For jobs that involve skills and tasks that are heavily impacted by AI and automation, we predict more job loss,” Gownder explains. “In job categories that are less impacted, obviously, we would predict less.” Forrester analysed 800 different job types. It seems that Art therapy is the right way to go. The Solow Paradox and AI Productivity Robert Solow’s famous observation that “we see computers everywhere except in the productivity statistics” finds a new iteration in the AI era. The parallel is instructive. It took nearly three decades for the internet’s productivity impact to materialise. E-commerce is only now truly disrupting traditional retail, as evidenced by the shuttering of independent shops from New York to Paris. Could Forrester’s five-year window be too narrow? Gownder acknowledges the limitation inherent in forecasting: “Anything that you forecast beyond five years is effectively an impression.” Yet the pace of technology adoption has accelerated dramatically. The telephone required 75 years to reach 100 million users from its 1878 introduction. The personal computer achieved the same milestone in 16 years. Mobile phones took seven years. ChatGPT? Two months. This compression suggests that while the Solow paradox may still apply, its timeline could be considerably shorter. “If there’s a job apocalypse, you’re going to have fewer people working because that’s what the apocalypse means. Those people would have to be producing more output. You cannot see a job apocalypse without aggregate productivity going up.” — J.P. Gownder, Forrester The productivity data tells a sobering story. From 1947 to 1973, US labour productivity grew at 2.7% annually. The current business cycle shows 1.8%. Even isolating the quarters since ChatGPT’s release yields only 2.2%. The numbers don’t lie, and they’re not yet showing the revolutionary gains AI proponents promise. Where the AI Job Impact Pressure Points Lie The AI job impact in the US will not be evenly distributed. Contact centre workers face continued pressure from automation that began with interactive voice response systems and now benefits from far more sophisticated solutions. Technical writers and web content creators occupy vulnerable ground. Insurance underwriters are seeing algorithmic encroachment; computer vision can now assess car accident damage from uploaded photos. Junior-level roles involving spreadsheet or presentation creation face mounting pressure. Software development presents a nuanced case. “If you are a junior level software developer,” Gownder notes, “we know that Claude does a great job of creating basic code.” Yet senior developers with architectural judgement and system-level understanding remain essential. The pattern repeats across knowledge work: AI augments more than it replaces, transforming job descriptions rather than eliminating positions entirely. “It’s not that there aren’t jobs that will go away,” he clarifies, “but they are much more specific and limited, and they need to be architected with the right technology to replace that job. It’s not everybody goes away.” Blue-collar work presents its own dynamics. Physical robotics will play a role in certain sectors: warehouse sorting and picking have improved through computer vision, and construction has seen experiments with brick-laying and cement-pouring robots. But the humanoid robots capturing media attention are unlikely to achieve significant workplace deployment within the forecast period. The physical world, with its infinite variations and unexpected challenges, remains stubbornly resistant to automation. The White-Collar AI Job Impact Misconception White-collar workers now constitute roughly 60% of the workforce in both the US and Europe, a dramatic shift from previous generations. These “symbolic analysts,” as Charles Handy termed them, don’t produce physical goods, which has led some to assume their work is easily transferable to AI systems. Gownder pushes back against this notion. “Most white-collar work is, in fact, fairly productive because there is something on the other end that someone is willing to pay for.” Software engineers create applications that enable other work. Physicians produce healthcare outcomes. Analysts help organisations make better decisions. The practical challenges of AI deployment in white-collar settings corroborate these theoretical objections. Hallucinations remain a persistent problem, introducing error margins that knowledge workers must catch and correct. Employees often lack the skills and understanding to use AI tools effectively. Organisations overextend their expectations of what AI can accomplish. “When it fails, it’s dramatic,” Gownder observes. The Deloitte incidents in Australia and Canada, where AI-generated content with obvious hallucinations reached government clients, illustrate the reputational risks of premature automation. The Australian government report contained fabricated academic citations and even a made-up quote from a federal court judgement. Both governments required refunds. “You don’t want to produce AI work slop and present it as your work without editing, without perspective. That is a losing proposition.” — J.P. Gownder, Forrester A Harvard Business Review study reinforces these concerns. Researchers found that executives who used ChatGPT to make predictions became significantly more optimistic, confident, and produced worse forecasts than those who consulted with peers. The authoritative voice of AI produces a strong sense of assurance, unchecked by the social regulation and useful scepticism that human consultation provides. AI Job Impact on Marketers and Digital Professionals For students entering digital marketing and related fields, the picture is complex but not nece

AI is not a tool, or is it? Reports regarding the impact of AI on jobs, society and businesses are cropping up all over the place at the moment in all corners of the world. Some of these reports are announcing forthcoming revolutions both for societies and our economies whereas others are playing down the impact of artificial intelligence, and reviving the good old Solow aka Productivity paradox (“You can see the computer age everywhere but in the productivity statistics”. follow up here and here). As a consequence, it is very hard to make an opinion, let alone advise business people and students alike with regard to what needs to be done in the future. Visionary Marketing has embarked on a mission to try and shed light on this topic in as rational and informed a way as possible. AI is not a tool, or is it? Should AI platforms become tawpayers? The great love affair of French people for taxes will not spare Artificial Intelligence Cavazza surmises. Indeed, according to him, AI is not a tool! A lot of these predictions are guided by ideology. The authors, be they proponents or opponents of AI, have a personal agenda, often political or ideological, and are trying to make facts stick to this agenda. This is not very useful. But others are based on fact and careful analysis. I have decided to focus on two of these reports/predictions. The first one is Fred Cavazza’s analysis of the impact of AI on society and the economy (original post in French), which describes Artificial Intelligence as a source of profound disruption. I have known Fred for years, and I know his deep knowledge of both subjects, which makes his report particularly valuable. With his kind permission, I have translated his piece from French to shed light on this subject. The other report is by Forrester’s JP Gownder, whom I’ll be interviewing soon. I will test Fred’s assumptions on JP and see what he has to say about this idea of disruption by AI. Hopefully, our readers, and especially my students who have a lot of pending questions about this, will be able to separate the wheat from the chaff after these two interviews and podcasts. AI is not a tool, it’s reshaping our society and economy AI can’t be seen as just another technological innovation. By establishing itself as a major driver of productivity, automation and decision-making, it’s fundamentally disrupting the economic and social balance of our society. Whilst the productivity gains brought by AI are already transforming office jobs and creating a chasm between employees who’ve embraced it and those who haven’t, a fundamental question emerges: how do we integrate these synthetic entities into our collective organisations? Between appropriate taxation, legal personality and psychological resistance, there are numerous questions to debate before we can draft a new social contract. AI IS NOT A TOOL — TLDR AI is triggering a disruption of our civilisation, it’s not just another tech breakthrough. It marks our genuine entry into the fourth industrial revolution by offloading, for the first time, human thinking and creativity to machines. AI’s productivity gains are already real and deeply uneven. A growing divide is opening up between workers who can work alongside AI and those stuck with 20th-century methods. AI agents are challenging how white-collar workers create value. Intelligent agents are transforming knowledge work, undermining certain business models and setting the stage for a rapid reshaping of office jobs. Integrating AI requires a new legal and fiscal framework. Like corporate entities, AI agents must be given a status that clarifies their responsibilities and reintegrates their value into the social contract. The socio-economic impacts reach far beyond just employment. AI affects our psychology, culture and demographics, making public debate crucial to head off looming social tensions. AI on the Davos Agenda This week, the world’s leaders are gathered at the Davos Economic Forum, and ecology isn’t on the agenda: AI, Big Tech and Trump Shine Most Brightly at the Davos Show . At Davos, the AI is not a toll debate was all the rage. Cavazza thinks that artificial intelligence will be a major disruptor not just of our exonomies but our societies too. AI is dominating every conversation, with considerations that extend far beyond technology: AI Is Poised to Take Over Language, Law and Religion, Historian Yuval Noah Harari Warns Palantir CEO says AI to make large-scale immigration obsolete “Artificial intelligence will displace so many jobs that it will eliminate the need for mass immigration” I’m not going to wade into commenting on everyone’s pronouncements, with their more or less biased viewpoints, but what’s certain is that major upheavals are on the horizon: AI and the Next Economy Nearly 80% of people feel unprepared to find a job in 2026 The AI revolution is here. Will the economy survive the transition? AI specialists are naturally the star guests at this 2026 edition of the Davos forum, invited to give their testimony and views: Deepmind and Anthropic CEOs expect AI to hit entry-level jobs and internships in 2026. Looking at it this way, it seems absurd to sit back as spectators whilst the AI revolution unfolds and do nothing to limit the fallout from this productivity shock. But not all’s lost—at least not for everyone, as countries in the global south are already gearing up for it: The AI Revolution Needs Plumbers After All. Productivity gains to be nuanced, but certainly not ignored I’ve had plenty of chances to explain generative AI’s impact (Superintelligence will multiply our capacity to act tenfold and The digital divide is a problem no one can ignore). Whilst we’re largely in agreement about what widespread generative models mean, there’s serious disagreement over the timeline for AI’s arrival. The dominant narrative keeps insisting that general AI is a pipe dream and that human intelligence is and will remain superior to machines. What is intelligence? This is precisely where ambiguities crop up: firstly, intelligence comes in many forms (Theory of multiple intelligences and What’s your intelligence type?); secondly, not all office work requires emotional or social intelligence. What I’m getting at is that most service sector jobs boil down to shuffling information and data between systems. You don’t need to be a genius to do that—AI can handle it with ease. To properly grasp the speed at which latest-generation AIs will gradually transform office jobs, I recommend you peruse the latest edition of Claude’s publisher’s macroeconomic barometer: Anthropic Economic Index 2026. Anthropic’s economis index 2026 For this fourth edition, the study’s authors analysed thousands of people’s activities using increasingly precise indicators: New building blocks for understanding AI use. This study yields several findings that demonstrate a strong progression in the adoption and capabilities of generative models. Notably, they observe an average 30% growth in Claude usage, driven mainly by the API rather than the chatbot—a sign of rapid adoption by advanced users (e.g., IT professionals) and slower uptake by ordinary users (white-collar workers using the web version). AI is not (just) a tool. As a matter of fact it’s not a tool at all, it’s a meta tool, a tool you can use to make tools.. The haves and the have nots A gap is therefore widening between those who’ve adopted new habits (working in tandem with AI) and those still working as they did in the 20th century. This gap is starting to become problematic, because the latest version of Claude (Opus 4.5) has capabilities comparable to those of an adult who’s benefited from over 14 years of education—the equivalent of a Bachelor’s degree. AI is not a tool but Clause isn’t a PHD either… yet. The question therefore is: how much longer can an employer justify paying salaries or hiring young graduates when chunks of the work can be farmed out to an AI? Whilst average productivity gains remain modest (1.8% according to the latest figures), AI’s contribution to certain tasks is absolutely spectacular: an average of 14 minutes to write a long article, versus 3 hours without AI assistance; an average of 5 minutes to analyse a complex data table, versus 1 hour 45 minutes without AI assistance. AI is not a tool, there are alo APIs You might argue this data’s skewed because these spectacular scores come from employees who are whizzes at using AI (therefore logically hyper-performers), but that’s not the case—the study covers ordinary employees with a 67% success rate for outsourced tasks. What this boils down to is that for a third of tasks, AI slashes processing time by 10 to 20 times in two-thirds of cases. If we apply some basic maths, AI can potentially triple efficiency—or to put it another way, cut the average time needed to complete a task by two-thirds. Which type of profile do you reckon managers will favour? (hint: McKinsey challenges graduates to use AI chatbot in recruitment overhaul) Soon the arrival of agentic white-collar workers Let me be clear: the productivity gains mentioned above relate to advanced AI usage, not just running searches in ChatGPT or asking Copilot to knock up meeting minutes. We’re talking about using generative models to their full potential, particularly intelligent agents (see Agentic Web: the revolution that won’t wait for you). Intelligent agents We’ve been banging on about these famous intelligent agents for a while now, but their potential only recently became blindingly obvious to ordinary employees (non-IT types) with the release of Claude Cowork, a very concrete wake-up call to the power o

Private equity branding remains one of the most underestimated levers for value creation in the investment world. While PE firms excel at identifying promising companies and optimising their financial structures, branding is frequently treated as an afterthought, reduced to logos and colour palettes rather than strategic assets. Yet the evidence suggests otherwise: strategic brand investment can dramatically shift market perception and, ultimately, company valuation. Marc Rust, Creative Director and Brand Strategist at Consequently Creative, has spent years demonstrating that branding deserves a seat at the strategy table. His striking claim that he transformed an $80 million company to look like a $120 million company through branding alone captures the essence of what strategic messaging can achieve when properly deployed. How Private Equity Branding Is Transforming Company Valuation With Storytelling The term “branding” itself creates immediate problems in private equity settings. At networking events, Rust finds that mentioning branding triggers what he calls “cognitive disruption” Beyond Logos: Redefining What Branding Actually Means The term “branding” itself creates immediate problems in professional settings. At networking events, Rust finds that mentioning branding triggers what he calls “cognitive disruption” – people immediately think of visual identity work that seems irrelevant to serious investment activities. Many professionals lack any clear definition of what branding encompasses, while others dismiss it as superficial design work. This misconception misses the fundamental truth: branding and messaging represent a powerful force for business growth that should inform strategy from the outset, not be bolted on afterwards as a cosmetic exercise. The real definition of branding, Rust argues, is “what you stand for in the minds of the people that you’re trying to reach, convert, and move into action.” This is not something companies own outright; rather, it is something they can influence through deliberate effort and sustained investment. The critical distinction lies between what companies do and why it matters. Most organisations focus their communications on deliverables and capabilities. Yet answering the question of why it matters opens doors to deeper insight about audience pain points, goals, and outcomes. This shift acknowledges that messaging exists not for the company but for its buyers, requiring communication in their language rather than internal jargon. The Evolution of Private Equity Strategy The private equity landscape has fundamentally changed over the past decade. The old-school approach – acquiring a company, trimming the fat, making it lean and mean, then finding a suitable buyer – no longer resonates with contemporary markets or the talent those markets require. Successful PE firms have embraced a different philosophy: nurturing acquired companies, building genuine value over time, and then pursuing exit strategies that reflect accumulated worth. This evolution makes branding more important than ever because value creation depends on perception as much as operational reality. When thinking about branding in private Equity, most people immediately think of visual identity work. All that seems irrelevant to serious investment activities even though it’s blatantly wrong, Mac Rust believes. Visual made with Midjourney Effective branding requires understanding multiple audiences simultaneously. Internal alignment comes first – the people who build products and deliver services need clarity about what their company stands for, especially during periods of transition. Post-acquisition, this alignment frequently suffers as employees wonder about new leadership, potential job losses, and strategic direction. Consequently Creative addresses this turbulence by bringing teams together to celebrate what they stand for, building stories around acquisition rationale and forward-looking plans grounded in existing strengths rather than imposed transformations. Beyond internal audiences, companies must establish clear market positioning relative to competitors and ecosystem partners. Finally, there are the buyers who will drive revenue growth during the holding period and, ultimately, the acquiring company that represents the exit opportunity. Each audience requires thoughtful attention, and branding provides the framework for addressing all of them coherently while maintaining a consistent core narrative. The Valuation Premium of Strong Brands Buyers demonstrably pay premiums for assets with strong brand equity. Companies that look more upscale and feel right command higher prices regardless of sector. This premium extends across every touchpoint: market presence, customer service quality, sales process sophistication, product presentation, and how offerings are described and positioned. The key lies in making everything about the audience – answering why customers should care and how specific features apply to their particular situations. Buyers demonstrably pay premiums for assets with strong brand equity, Rust declares. Visual made with Midjourney Building a brand encompasses far more than marketing communications. Yet smaller companies actually hold advantages here that larger organisations lack. Without established brand perceptions moulded into market consciousness over decades, mid-market companies enjoy flexibility that industry giants cannot match. They can position themselves as something new even when their offerings are not particularly novel, or emphasise technology, audience needs, or other differentiating angles. The argument that mid-market companies lack resources for serious branding investment misses this opportunity – budget allocation to branding should be generous precisely because returns can be substantial and the competitive playing field favours agility over scale. AI as Tool, Not Solution The artificial intelligence revolution has created new temptations for companies seeking branding shortcuts. Tools now generate logos, mission statements, and complete brand architectures almost instantly. But Rust cautions strongly against treating AI as a solution rather than what it actually is: a technology that should come last in any strategic process. The POST method he advocates begins with understanding people (your audience), then defining objectives (business goals), followed by strategy (how to achieve those goals), and only then selecting technology. Flipping this sequence – jumping on AI because everyone else has it – represents precisely the wrong approach to brand development. The danger of AI-driven branding lies in acceptance without scrutiny. When tools generate content quickly, users become passive recipients rather than active directors, keeping their eyes closed and allowing technology into the driver’s seat. Rust draws on singer-songwriter Tom Waits: “The world is a hellish place and bad writing is destroying the quality of our suffering.” AI contributes to this problem when deployed thoughtlessly, generating content that lacks the provocative point of view necessary to differentiate companies in crowded markets. Bad content existed before AI, but artificial intelligence is intensifying the problem. The world is a hellish place and bad writing is destroying the quality of our suffering Tom Waits That said, AI offers genuine utility when approached correctly. Brainstorming, idea generation, concept testing, and data synthesis all benefit from AI assistance. The technology serves well as a sounding board for strategic thinking. The crucial distinction is maintaining human agency – staying in the driver’s seat rather than ceding control to automated systems that cannot understand business context or competitive dynamics. B2B Private Equity Branding: The Relationship Imperative The notion that B2B companies need branding less than consumer-facing businesses deserves serious challenge. Branding fundamentally concerns relationship-building, and relationships involve humans making decisions regardless of whether they represent individual consumers or institutional buyers. When someone purchases at a supermarket, they often choose the best-looking product rather than the one with objectively superior ingredients. B2B purchasing follows similar patterns – everyone wants to work with companies that appear capable, innovative, and aligned with their values. “The notion that B2B companies need branding less than consumer-facing businesses deserves serious challenge” B2B branding may require less ongoing investment than B2C equivalents because it depends less on constant social media presence and retargeting campaigns. However, the fundamental mechanics remain identical: building trust through consistent value delivery over time. Each interaction with a company should provide something useful, and these value contributions compound into trust. Value + value + value = trust – a formula that applies regardless of whether customers are individuals or organisations. The Research Imperative: Discovering Hidden Stories The biggest mistake private equity firms make when rebranding after acquisition is proceeding without empathy for audiences. This criticism is not meant to disparage PE professionals – it simply reflects that branding expertise lies outside their core competencies. The solution involves partnering with agencies that understand how empathy drives both growth and culture. Jumping straight to visual refresh without strategic groundwork means missing reasons to believe that proper research would uncover. Rust illustrates this with two compelling examples. Working with a company owning approximately 100 senior living properties across the United States, his

The ebook self-publishing landscape has undergone a remarkable transformation over the past decade. What was once viewed with scepticism by the publishing industry has become a legitimate and often preferred path for authors worldwide. To understand the current state of this evolving market, we spoke with Kris Austin, whose platform Draft2Digital serves over 300,000 authors publishing more than a million titles across global markets. From Oklahoma City, he shared his insights on how independent authors are reshaping the publishing world. Inside the Ebook Self-Publishing Industry With market shares amounting to 40% of sales in the US, ebooks present new opportunities for writers who are able to benefit from self-publishing platforms like Self2digital. Can you introduce Draft2Digital and its mission? Draft2Digital currently serves over 300,000 authors who are independently publishing more than a million titles. We have been operating since 2012, and the industry has changed considerably during that period. Our goal is to help authors achieve their dreams by removing technical barriers and making the publishing process as streamlined and straightforward as possible. What languages and markets do you cover? We have published books in over a hundred languages. While English remains predominant, approximately 15 to 20 percent of sales come from non-English titles, with Spanish and German ranking as the second and third most popular languages. Our distribution reaches 180 countries, and about 40 percent of all sales occur outside the United States. ebook self publishing industry entrepreneur Kris Austin talked to us from Oklahoma City, OK. How has self-publishing evolved since 2012? When we started in 2012, self-publishing was still in its early stages. The real catalyst came in 2007 when Amazon released the Kindle, which sparked the explosion of digital books. Back then, there was significant stigma attached to being an independent author; many felt they were not as credible as traditionally published writers. Today, that perception has completely shifted. Many authors now choose self-publishing as their first option. We also see numerous hybrid authors who move between traditional and independent publishing, depending on their goals. The focus has shifted to the quality of the book and reader demand rather than the publishing model itself. What types of books dominate the ebook market? The majority of our ebooks are genre fiction: romance, fantasy, mysteries, and thrillers. These narrative fiction categories account for approximately 80 percent of ebook sales. Our print-on-demand service shows a different pattern, with roughly 40 percent fiction and 60 percent non-fiction. All these books are intended for consumer readers purchasing for personal enjoyment. Genre fiction (romance, fantasy, mysteries, and thrillers) amounts to approximately 80 percent of ebook sales Wit ebook self-publishing, authors can find readers anywhere in the world without leaving their homes. Image created with Midjourney Is ebook self-publishing viable for image-heavy books like photography? It is possible, though more demanding. Image-heavy books typically require a professional formatter to achieve the desired layout, particularly in digital formats where presentation can be challenging. For print editions, colour printing and layout involve additional complexity compared to text-only publications. What determines success in ebook self-publishing? The most successful authors treat publishing as a business. After creating a book they are proud of, they focus on marketing, discoverability, sales, and distribution. They approach it with an entrepreneurial mindset. However, it can also work as a part-time endeavour, particularly for authors writing series with multiple titles. One advantage of independent publishing is that you do not need a massive readership to succeed. Indie authors typically retain 60 to 80 percent of their sales revenue, allowing them to price competitively and target niche markets effectively. Even with just 2,000 potential readers, if you capture that audience and build loyalty, you can build a sustainable career. Indie authors typically retain 60 to 80 percent of their sales revenue If writing is your dream, ebook self-publishing could make it real draft2digital claims. How does Draft2Digital help authors reach global audiences? First, availability is essential. Authors upload their manuscript in Word format to our website, along with a cover image. I recommend not spending more than 100 dollars on a cover when starting out. Our system converts everything to digital formats and distributes to thousands of stores, including major online retailers, smaller platforms, and libraries across the US, UK, and Australia, typically within a few days. Accurate metadata, including title, description, and category, is crucial for helping readers find your book. What marketing strategies work for unknown authors? Discoverability is always a challenge. Successful authors connect with readers through social media, choosing platforms based on their target audience. Facebook may suit an older demographic, while TikTok reaches younger readers. Authors must identify where their audience congregates and invest effort in building those connections. Nothing comes free when selling a product; it requires consistent work. What are the main differences in reading habits across countries? Reading preferences vary significantly by region. Some countries, like the US, have high ebook adoption, while others, such as Germany, still favour print by a considerable margin. Certain markets, like Canada, show preferences for book bundles. Interestingly, German readers consume many English-language books, so we sell substantial quantities of English print titles there. What is the current balance between ebooks and print? When ebooks began growing around 2007, there were widespread concerns about the death of print. That never materialised. Ebook growth peaked around 2013, but print remained dominant. Currently, approximately 60 percent of books sold are print and 40 percent are ebooks, though this varies by genre. Romance readers predominantly purchase ebooks due to lower cost and convenience, while non-fiction readers prefer print for its tactile qualities and ease of reference. This ratio has remained relatively stable for years. Approximately 60 percent of books sold are print and 40 percent are ebooks Are people reading less than before? Readership fluctuates in cycles. We saw a significant peak during the COVID lockdowns, and we have been coming down from that high. However, engagement appears to be recovering. Books now compete with digital streaming and social media for attention, but dedicated readers will always find their books. We are optimistic that younger generations will discover books that resonate with them and develop reading habits. How is artificial intelligence affecting the ebook market? AI-written books exist throughout the market. We support AI as a tool for outlining, brainstorming, and various other assistance, much like word processors and spell checkers became standard aids. What we do not support is fully AI-generated content. AI-written books have become a significant challenge This has become a significant challenge, with platforms like Amazon being flooded with such material. It harms the industry and makes it harder for readers to find quality books. While AI may eventually produce excellent literature, we are not there yet, and this remains an ongoing challenge for the market. How do you help authors stand out in a crowded marketplace? We maintain merchandising relationships with all major retailers. Our mission is to identify promising books and propose them for spotlight placement and promotional features. We submit thousands of titles annually and achieve a 60 percent success rate. Authors can apply through our website to participate in these programmes. We also invest heavily in author education through our Self Publishing Insiders podcast, where we interview industry leaders, successful indie authors, and service providers to help authors improve their marketing and sales strategies. What do you predict for the future of digital publishing? Independent authors have proven their agility and ability to respond quickly to reader demands in an industry historically slow to adapt. Traditional publishers are increasingly looking to indie authors for insights on how to operate differently. They are actively recruiting successful independent authors into the traditional world. I expect traditional publishing to adopt more characteristics of indie publishing: greater agility, flexibility, and responsiveness. This convergence will continue accelerating over the coming years. Final thoughts about ebook self-publishing The ebook self-publishing revolution has fundamentally altered the publishing landscape. What Kris Austin describes is not merely a shift in distribution channels but a democratisation of authorship itself. With platforms like Draft2Digital removing technical barriers and providing global reach, the determining factors for success have shifted from gatekeepers to readers. For aspiring authors, the message is clear: quality content, business acumen, and direct reader engagement now matter more than ever. The stigma of self-publishing has given way to recognition that, ultimately, a book’s value lies in its ability to find and satisfy its intended audience, regardless of how it reaches them. Learn more at Draft2Digital The post Inside the Ebook Self-Publishing Industry appeared first on Marketing and Innovation.

Whereas artificial intelligence is reinventing Web writing, the written word has never been more valuable. Selim Niederhoffer, a copywriting trainer and bestselling author, has recently been exploring how marketing professionals can still succeed amidst “enshitification“, online influence, and automation. Meet an expert who remains confident in the power of words. Copywriting in the age of AI: why words retain all their magic Niederhoffer is adamant: despite GenAI, the written word retains its magic — image produced with Midjourney Human vs. AI Selim, a seasoned copywriter and author is keen on using what he calls “magic words.” Ask him why and his answer might surprise you in this age of sheer automation. “I’ve already been thinking about magic words for years. I want to dig deeper, show real examples and, most importantly, explain why they actually work.” Pen and paper His approach remains resolutely traditional. “I genuinely work with a pen and paper.” This approach reveals something fundamental about Web writing: it works when you truly understand the psychology behind it, not when you just mechanically apply copywriting techniques. Selim bases his work on well-established principles of persuasion. Selim Niederhoffer still believes in the virtues of word magic in Web writing — image produced with Midjourney “When there’s a principle of persuasion, there’s usually a word that goes with it. Take urgency or scarcity, for example. If something’s rare, that means limited places, running out of stock… that sort of thing. That’s how I build my word cluster. The research then extends to field observation. “I look at what my clients are using, what’s going on at Burger King, McDonald’s, Nike. I check out the major brands too – what they’re doing on YouTube, on LinkedIn.” Eventually, Selim identified 55 magic words but trimmed them down to 50. It’s an approach that perfectly shows what still sets humans apart from machines: the ability to critically analyse and curate with discrimination. However, it’s worth adding nuance. Selim can’t conceal that he genuinely “loves” ChatGPT. As we’ll see later, this raises some legitimate questions. Thank you! the ultimate magic word Among the 50 words he’s analysed, the first is also the simplest: thank you! For Selim, this should be essential for every business. “How many times have you walked out of a shop where the sales assistant just didn’t seem to have noticed you? Whether you bought anything, or didn’t doesn’t make any difference, once the transaction’s done, you’re out the door.” Yet some brands know how to thank their customers. “Nespresso or Apple, for example: the Nespresso employee comes out from behind her counter, she hands you your product. Thank you for your visit. Have a nice day! That’s how it should be.” For Selim, saying “thank you!” is more than just being polite, it’s a way of life. “You can go further: thank you for your visit, thank you for subscribing to the newsletter, thank you for your comment. You need to constantly think in terms of gratitude.” This approach fits into what Gary Vaynerchuk called The Thank You Economy. “We’re in an attention economy,” Selim explains. The stakes are high in Web writing: how do you maintain this human dimension at corporate scale? “For me, the essence of business is that there’s a person in front of you who’s exchanging something with you. That’s really the foundation. But today, how do you keep that at corporate scale?” Data confirms an intuition: gratitude improves customer experience and encourages loyalty. A valuable lesson for all those who practice Web writing and seek to create a lasting connection with their audiences. You need to seek to create a connection with your audiences — image produced with Midjourney AI and Web writing: threat or opportunity? The conversation inevitably turns to artificial intelligence. For Selim, AI is first and foremost an incredible productivity tool. “If I’ve got a newsletter to write, I’ll use AI. Sales pages? AI. The thing is, AI works brilliantly for me. I’ve even become clearer in my writing,” he admits without hesitation. But he’s not entirely starry-eyed about it. He’s identified three major pitfalls that professionals absolutely must avoid. The three pitfalls of AI in copywriting The first pitfall is wordiness. “ChatGPT tends to waffle because it’s been trained on Reddit, Wikipedia, and similar sources. However, in real life, when we write, we cut down to the core, deleting most of what we initially write.” The second limitation is related to syntax. “AI tools have their favourite phrases. It can be a bit clunky. It’s got that ‘ChatGPT-ish’ quality – you know it when you see it. After a while, you can spot AI-written text a mile off.” The third issue, and the most surreptitious one, is lack of personality. “When we just use basic LLMs, we lose our tone of voice, we lose what makes us different, we lose what makes people go ‘Ah! That’s Selim, that’s Yann’. That personal touch. I’d say that’s the biggest danger AI poses for copywriters, Web writers, anyone working in this field.” For him, the key questions are: how do you refine AI, how do you avoid its main pitfalls, how do you stay in control and how do you harness all its power? This evolution shows a profound shift in the profession: the copywriter is becoming an orchestrator of AI platforms. Losing our skills Selim warns against a hidden danger: declining skills. He reminds us that the brain is a muscle, and using it creates connections. However, if we stop using it, those connections are lost. He admits that he has fallen victim to this himself. “I have noticed that my writing isn’t as fluid as it used to be when I’m starting from scratch. Between 2010 and 2022, I was churning out three to five blog posts a week. Now, if I can just knock up a prompt, get a result and tweak it a bit, job done. But it’s less satisfying.” This awareness led him to experiment. “I run A/B tests. Send out version A written by AI, version B written by me, then see who clicks more. I check which headlines work best, which text performs better,” he explains. It’s a data-driven approach that could bruise his ego, he admits with a laugh, but it’s essential for understanding what genuine human value looks like in Web writing. What human added value tomorrow? The final question comes naturally: will humans still add value compared to machines in a year or two? Selim remains cautious about making predictions. “I’m rubbish at forecasting because I don’t spend enough time on it,” he admits candidly. But his reading of the market reassures him. While I can see companies increasingly launching AI training for their staff, there’s always the question of depth. Are they adequately trained and properly supported? Web writing needs remain massive “I meet marketing directors who want me on their pay-roll. We’ve got so much work to get through that we’d need ten people using ChatGPT just to keep up,” he reports. This suggests that AI isn’t so much replacing writers as it is multiplying opportunities for content production. Selim’s thoughts tap into bigger questions about how work is evolving. “The idea of heading towards a society where people become useless doesn’t surprise me. I’ve seen it enough times in sci-fi,” he confides. This vision actually mirrors history: villages like ours in the Pyrenees, with thousands of inhabitants at the turn of the 20th century now down to a few dozens. Economic and technological shifts have always redrawn the employment map. What the future holds for Web writing This conversation with Selim Niederhoffer sketches out the shape of Web writing as it transforms. Words remain magical, but how we summon that magic is changing. AI’s becoming a powerful content production tool, but the real value still lies in properly understanding how persuasion works, in being able to orchestrate these tools cleverly, and in keeping a human tone of voice. Here’s the paradox: just when machines can churn out infinite text, what becomes scarce is quality strategic thinking, subtle psychological understanding, and the ability to inject real personality into content. Selim’s magic words are just a a beginning – they are mere ways of understanding what makes human readers tick. A hybrid Web According to Selim, tomorrow’s Web writing will be hybrid – a collaboration between human and machine. It will demand technical mastery of AI tools, but it’ll be rooted in a deep understanding of how persuasion actually works. In short, words will always be magical, but that magic needs cultivating, training, constant work – otherwise it will erode in the face of automation’s easy appeal. One needs to keep reading, discussing and engaging with real people. That’s how we will train our brains and hang onto that human edge that makes all the difference in Web writing. 10 of Selim’s 50 magic words Here are 10 words taken from Selim Niederhoffer’s book “Les Mots Magiques” (Magic Words), which are effective words to boost your marketing content and encourage action: Please Thank you New Since Out of stock Free Now or never Satisfaction guaranteed or your money back Hundreds of people already trust us First name The post Web writing : words retain all their magic appeared first on Marketing and Innovation.

Commentary on the environmental impact of AI often swings wildly between doom-and-gloom catastrophism and blind techno-optimism. But where’s the truth in all this? On July 24, 2025—the symbolic date of Earth Overshoot Day—we sat down with Yves Grandmontagne, founder and editor-in-chief of DCMAG (Data Centre Magazine*), to get his take on AI and its real environmental impact. It is worthy of note that Yves and I explored Silicon Valley’s infrastructure innovators together through extensive press tours some time ago. This provided us with firsthand insight into the tech industry’s approach to these challenges. The current annotated transcript of our interview is a summary of our thorough, nuanced, and let’s admit it, quite lengthy discussion. You are therefore encouraged to treat this article as your starting point for diving deeper into this extremely complex topic. Exploring the Real Environmental Impact of AI What’s the real environmental impact of AI? An employee keeps watch over the cooling units at Orange’s data centre in Val de Rueil in Normandy, France — Photo antimuseum.com * DCMag is only available in French This post summarises what turned out to be an incredibly rich hour-long conversation. The sheer complexity of this topic forced us to dig into multiple technical, economic, and environmental angles—making any kind of comprehensive analysis near inconceivable. Drawing on his deep expertise in the data centre and AI sectors, Yves Grandmontagne gives us some much-needed factual perspective on a debate that’s often polarised between doomsday scenarios and over-the-top techno-optimism. To tackle this properly, we decided to take recent quotes—both positive and negative—and fact-check them with our expert. Yves’s analysis helps us cut through the noise and understand what’s really at stake in this technological breakthrough. Environmental Impact of AI: Reality Check Time TLDR: Environmental Impact of AI The electricity consumption issue is more nuanced than you think* – AI will represent 20-30% of data centre consumption (not twice that number), and only 2-4% of overall electricity consumption Energy efficiency gains are actually remarkable – Over the last decade: number of data centres x2, floor space x4, but energy consumption up only 6% Beware of dubious comparisons – Comparing a ChatGPT query to Google search is methodologically flawed (completely different technologies and services) Water consumption varies massively by geography – Huge issue in the US, but Europe has been using smarter closed-loop systems for ages Tech innovations look promising – New technologies (direct liquid cooling, immersion cooling) are slashing water and energy consumption AI might actually be part of the solution – Can optimise energy mix management and electricity transport, which is currently our main bottleneck Let’s get some perspective here – Data centre impact remains pretty marginal compared to the chemical industry (32% of French energy consumption) or agriculture *All numbers by Yves Grandmontagne at Data Centre Magazine Bottom line: The impact is “real but massively overstated”—we need to put things in context and remain cool and collected. So here are the quotes about the environmental impact of AI that we wanted to fact-check with Yves. Rumours vs Reality: Decoding the Doomsday Predictions Predictions of Booming Electricity Consumption The first claim I put to Yves Grandmontagne was Synth Media’s prediction [Fr] that “AI’s growth could double data centre electricity consumption by 2026.” His response immediately throws cold water on this alarmist take: “It’s absolutely true that AI is rolling out infrastructure at breakneck speed and eating up more space in data centres. Sure, it’s going to significantly bump up their energy consumption—no question about that. But will consumption actually double? I seriously doubt it. AI should account for somewhere between 20 and 30% of global data centre consumption worldwide.” A data centre server rack — photo antimuseum.com AI should account for somewhere between 20 and 30% of global data centre consumption worldwide This reality check reveals something consistent throughout Yves Grandmontagne’s analysis: the absolute need to put numbers in context. The expert points out that this increase is just part of the natural progression tied to our ever-growing digital habits. “What’s driving this consumption increase is our daily usage, whether that’s for work or for personal reasons,” he reminds us, highlighting our collective responsibility in this evolution. It’s a topic we’ve tackled before with a broader focus on digital consumption. The most striking part of his analysis is about the energy efficiency angle. Contrary to popular belief, data centres aren’t following a consumption curve that mirrors data growth. This improving efficiency is something that gets completely overlooked in public debates about the environmental impact of AI. ChatGPT’s Carbon Footprint: More Context Needed When it comes to the 10,113 tonnes of CO2 equivalent attributed to ChatGPT usage in January 2023 (Basta Media – Data for Good, AI has the potential to destroy the planet), Yves Grandmontagne takes a refreshingly pragmatic approach: “I can’t verify that exact figure. Getting that precise—down to 10,113 tonnes—represents a massive methodological challenge, especially when you’re dealing with AI infrastructures that are distributed systems.” We asked Yves Grandmontagne, editor-in-chief of DCMAG (Data Centre Magazine) to give us the real story about the environmental impact of AI — He gave us facts and figures, which we’ve compiled in the infographic at the end of this post This observation raises a crucial methodological point: just how tough it is to accurately measure the carbon footprint of distributed infrastructures. The expert does acknowledge this pollution is real, but puts it in perspective: “Those 10,113 tonnes of CO2 still represent volumes significantly smaller than what many other industries pump out.” This contextualisation isn’t about downplaying the issue—it’s about keeping things proportional. Yves Grandmontagne reminds us of a basic truth that is often overlooked: “The moment we use our smartphones, we become CO2 producers.” This highlights the inconsistency in criticisms that single out AI from our overall digital consumption. The Google vs ChatGPT Comparison: A Methodological Trap Watch out for convenient shortcuts between pollution and digital tech—they’re everywhere and pretty handy when you want to hide overall industrial pollution — image created with Midjourney MIT’s claim that “a ChatGPT query uses ten times more electricity than a Google search [p 9]” perfectly illustrates the danger of oversimplified comparisons, according to Yves Grandmontagne. His take is particularly eye-opening: “This comparison just doesn’t work. I think we’re making a fundamental methodological error here because we’re comparing two completely different technologies. Google is a search engine that gives you results that you then have to sift through to find what you’re actually looking for. ChatGPT, on the other hand, serves up information that’s already structured and ready to use.” “The more we use computing, and the more we use AI, the more we consume.” User responsibility matters — images antimuseum.com Orange data centre in Val de Rueil This analysis shows just how sophisticated you need to be when evaluating the environmental impact of AI. A ChatGPT query might actually replace multiple Google searches plus visits to various websites. That makes direct comparison pretty meaningless. This distinction will probably become moot anyway with the rollout of Google AI Overviews, which will integrate similar functionality to ChatGPT. The Water Issue: Geography Matters More Than You Think New data centre cooling techniques work in closed loops. Image antimuseum.com Orange data centre in Val de Rueil Massive Geographical Differences One of the most revealing parts of our interview dealt with data centre water consumption. Yves Grandmontagne draws a crucial distinction between American and European practices: In the United States, when you install a data centre, you’re a private company reaching out to other private companies for water on one side, electricity on the other. And the utility companies that manage energy or water are thrilled to have a massive client that’ll absorb a chunk of their production. This geopolitical insight explains those 5.4 million litres of fresh water attributed to ChatGPT-3 training. The issue isn’t the technology itself—it’s local practices and regulations. In Europe, our expert reminds us, “we’ve been developing infrastructure cooling systems that work in closed circuits or are air-based rather than water-based for quite some time”. Take Orange’s data centre in Val de Rueil, for instance—it’s cooled by the crisp Normandy air. The only exceptions are during heat waves, which are naturally time-limited. How Cooling Actually Works Yves Grandmontagne’s technical explanation demystifies the cooling process: “Water acts as a conductor to capture heat.” He then breaks down the dual-circuit system that protects infrastructure while managing thermal exchanges. This approach reveals that warm water discharge from data centres (~20-25°C) stays well below that of nuclear plants (27-35°C for the Gravelines nuclear plant in the North of France). That gives us a useful comparison point for our debate. Warm water discharge from data centres (around 20-25 degrees) stays well below that of nuclear plants (27-35°C for G

Will the AI bubble burst or is GenAI here to stay? The artificial intelligence industry is experiencing unprecedented financial euphoria. Yet, the current situation is very confusing. AI investments are reaching dizzying heights. Let’s mention OpenAI’s $40 billion funding round at $300 billion valuation and Mistral AI’s €1.7 billion funding round. Yet, some commentators are very critical of the situation. For instance, Ed Zitron predicts that the AI bubble will burst in Q4 2025. All this is fueling intense, rather than rational, debate. I wanted to confront these concerns with the expertise of Bernhard Schaffrik, Principal Analyst at Forrester Research. His analysis is insightful and nuanced. In his mind, there will be some sort of correction, but at the same time, GenAI is too popular to disappear. When Will the AI Bubble Burst? Is the AI bubble about to burst or is GenAI here to stay? Forrester’s Schaffrik predicts corrections but says GenAI is too popular to go — photo by Forrester.com Forrester’s Bernhard Schaffrik is recognized as one of the most insightful experts in artificial intelligence. He provides a nuanced analysis that transcends simple financial considerations. His perspective on the AI bubble burst scenario offers first-hand insights for understanding where this transformative technology is truly heading. The AI Bubble: Financial Reality, Technological Continuity The question of a potential AI bubble burst cannot receive a univocal answer. As Bernhard Schaffrik rightfully points out, it all depends on one’s perspective. This duality of vision probably constitutes one of the keys to understanding the current situation and the likelihood of an AI bubble burst. Like us, Schaffrik righfully points out that the main issue with AI is societal and philosophical — image generated with Adobe Firefly “It’s almost impossible to get a one-sentence response from an analyst. Allow me two sentences. Number one is, of course, it always depends on the role or the profile you’re asking. If we are talking about financial investors, then yes, there are strong signals of this being a bubble because there is so much money being pumped into it—more than $120 billion US dollars in capital expenditure on AI infrastructure alone, just by the Magnificent Seven tech providers. So that bubble could burst,” explains Forrester’s expert. This assessment gains particular relevance when considering Google’s $9 billion AI data center investment in Oklahoma for advanced AI training infrastructure. This financial perspective, however, tells only part of the story. Technological adoption follows a different logic from financial markets, as Schaffrik confirmed during our exchange about the AI bubble burst potential. “But now, if you put yourself in the shoes of enterprise decision makers, tech decision makers, also AI users, there are many who would say, ‘I don’t care if that bubble bursts, the technology is there, and it won’t go away.’ “Regardless of the amounts all the financial transactions surrounding the AI industry, people are actually using this technology. And they like what they are seeing. It might not be the disruptive, transformative value some are surmising. It’s probably more incremental than that, but the adoption of that technology is undeniable.” The Revenue Challenge: A $25 Billion Gap to Bridge Fortune’s analysis reveals a concerning gap between current investments and generated revenues. To justify current investments, AI companies would need to generate $40 billion in annual revenue, while they currently produce only $15 to $20 billion. Schaffrik doesn’t believe in an AI bubble burst right now — image made with Adobe Firefly I was wondering whether this $20-25 billion gap could represent a systemic risk that could trigger an AI bubble burst. Schaffrik remains relatively optimistic on this point: “There is still enough money in that market to back these revenue gaps at least for a while. And what I’m also seeing is that especially when it comes to the largest enterpriseson the planet, they are convinced to continue using that software. And if it comes at a premium which is decent, arguably, maybe a couple percentage points higher than what they are paying today for the software, then this seems to be acceptable.” This acceptance of additional costs by large enterprises stems from the incremental value they perceive, even if it hasn’t yet reached the promised transformation level that might prevent an AI bubble burst scenario. LLM Regression: A Warning Signal? A particularly troubling element in the current ecosystem is the recent NewsGuard study revealing that major LLM systems are no longer progressing but regressing, generating more hallucinations and errors. This observation raises fundamental questions about current technology maturity and its impact on AI bubble burst predictions. “I’m not saying that LLMs and generative AI are progressing in a linear fashion nor that this technology will be disruptive in any way, despite the promise. As we have seen with emerging technologies for decades and even centuries, it takes breakthrough technological revolutions rather than evolutions to fulfill such promises,” analyzes Schaffrik. This vision of the current limitations of AI doesn’t diminish Bernhard’s long-term optimism: “But I’m also convinced that these breakthroughs will happen, not within the next seven, eight, nine, 12 months, but maybe in the long term. Something else will be coming up.” Energy Efficiency: The Achilles’ Heel of AI One of Schaffrik’s most compelling criticisms concerns the energy efficiency of current systems. His comparison between the human brain and data centers is striking and relevant to understanding whether we’re facing an AI bubble burst. “If we look at the amount of energy our brains are requiring to create a certain inference, and how much an LLM would require to achieve the same result with electricity, this cannot be the way forward.” This energy inefficiency constitutes a major barrier to scalability and will require significant technological breakthroughs to overcome, potentially influencing AI bubble burst timing. Pilot Failures: Business as Usual or Red Flag? The 95% failure rate of corporate AI pilots revealed by MIT research might seem alarming and suggestive of an impending AI bubble burst. Yet Schaffrik places this figure in its historical context: “It’s quite normal. As an analyst covering innovation management, what I have been observing over time is that about 10% of all innovation-related minimum viable products, proof of concepts, pilots, will turn into a product.” The problem would rather lie in unrealistic expectations: “Everybody rushed at it because one believed that since it’s accessible through natural language, it should  be easier to deploy, to implement, and there are no drawbacks and negatives. That might explain that the failure rate is slightly higher than with technologies we saw in the past.” This assessment aligns with Gartner’s prediction that 30% of GenAI projects will be abandoned after proof of concept by 2025 due to poor ROI and unclear business value. AGI: The Next Revolution in Motion Despite current limitations, Schaffrik maintains his bold prediction from his July 2025 analysis “Demystifying Artificial General Intelligence” that Artificial General Intelligence (AGI) represents “the biggest change in tech we have ever seen and is starting right now.” This vision, which could influence AI bubble burst scenarios, is structured around three maturity stages. “Competent artificial general intelligence is lurking around the corner. Our prediction is that between 2026 and 2030, we will see competent artificial general intelligence. You could think of it as your first trustworthy AI agent. You might not want to give it your car keys or your wallet, but it might do amazing things.” The subsequent stages would unfold over more distant horizons: independent AGI within the next five to ten years, And lastly, strategic AGI in a more distant future. Current AI Limitations: Experience versus Training A crucial point raised in our discussion concerns the difference between training and experience. As I pointed out to Schaffrik, experience develops critical thinking that current LLMs don’t yet possess, which could impact AI bubble burst predictions. “We might get to a point where most of us humans wouldn’t be able to tell if on the other side, a human or a machine is interacting with us. There will be areas where we will still be able to tell. But experience is something we could at least partially solve with more data and better data,” states Schaffrik. The solution, according to him, lies in massive data collection: “So much of the billions of investment money flowing now into all these big companies is also to collect and curate data also from the physical environment. Bringing all this data together, will create something that mimics experience.” The Human Factor: What’s Left for Us? The philosophical question of what humans will do when machines surpass us in thinking capabilities represents Schaffrik’s personal concern regarding potential societal implications of advanced AI, regardless of any AI bubble burst. “That’s my personal doomsday scenario, I must say. It’s not good for us humans to just idle around. So it’s not so much a technical conversation, but more a political, societal, psychological and philosophical one. So I’m sure we are far away from this, but we are getting there.” Leadership and Preparation: The Governance Challenge Regarding leadership in this transformation, Schaffrik acknowledges the complexity: “Rulers are supposed to rule