{"slug": "youre-not-adopting-ai-youre-paying-for-it", "title": "You’re not adopting AI. You’re paying for it.", "summary": "A new report from Trinity College Dublin and Microsoft warns that 92% of Irish organizations are using or planning to use AI, but the gap between firms getting real value and the rest is widening. Guest author Rahim Hirji argues that most AI rollouts fail because they bolt the technology onto unchanged processes rather than redesigning work around it, citing economist Paul David's 1990 paper on electrification as a historical parallel.", "body_md": "*Guest post by Rahim Hirji author of SuperSkills: The seven human skills for the age of AI,*\n\nSomewhere in your business there is probably an AI subscription that seemed like a good idea at the time. The pilot impressed everyone, a handful of people now use it daily, most tried it twice, and the transformation it was meant to deliver has settled into something closer to a modest convenience. If that sounds familiar, you’re in the majority. The AI Economy Ireland 2026 report from Trinity College Dublin and Microsoft found that 92 per cent of Irish organisations are using or planning to use AI, and warned in the same breath that the gap between the firms getting real value and the rest is widening.\n\n## The challenges of AI\n\nThe usual explanations blame the technology. The model wasn’t good enough, the vendor oversold it, the hype ran ahead of the reality. Sometimes that’s fair. More often the tool did roughly what it promised, and the disappointment came from somewhere older: nothing changed around it.\n\nWe have made this mistake before, at industrial scale, and an economist wrote the definitive account of it thirty-five years ago. When factories first electrified in the 1890s, most owners pulled out the steam engine and installed an electric motor in the same central position, driving the same overhead shafts and the same leather belts to the same machines in the same layout. A new power source, dropped into an unchanged factory. Paul David’s classic paper, The Dynamo and the Computer, explained why the expected productivity gains were delayed for decades. They arrived only in the 1920s, when a new generation of engineers stopped asking where to put the motor and started asking what a factory should look like now that power could go anywhere. Small motors on every machine, floors arranged around the flow of work instead of the reach of a drive shaft. Once the work was redesigned around the technology, electrification drove roughly half of all manufacturing productivity growth in that decade.\n\nEvery AI rollout since has faced the same fork, whether anyone in the room named it or not: bolt-on or redesign. The bolt-on swaps the power source and keeps the factory. The redesign asks what the work should look like now that this capability exists.\n\nIn the work itself, the bolt-on fails in a predictable way. When you automate one step of a process built around humans doing that step, you don’t remove the constraint, you move it. A team uses AI to produce first drafts in minutes instead of days, and the review stage, still staffed and structured for the old pace, becomes the new queue. Output rises at one point in the chain, piles up at the next, and the tool takes the blame for a process that was never re-plumbed to absorb what it now produces.\n\nIn people, the bolt-on fails more expensively. Handing staff AI access without developing their ability to direct it, judge it and check it isn’t adoption, it’s abdication. The capability that matters is not prompting, which anyone can pick up in an afternoon. The scarce skill is knowing when the output is wrong: the confident answer that doesn’t hold, the summary that flattened the one detail that mattered, the analysis that reads as plausible rather than true. Google Cloud’s DORA research put it plainly last year: successful AI adoption is “a systems problem, not a tools problem.” Most AI budgets still treat it as one.\n\nAnd even within the people, capability is never one number. It’s a spread. In the same firm, on the same tools, with the same access, I’ve watched people struggling with the basics and trying hard, people defaulting to the tool for a quick answer and stopping there, and a few whose workflows have rebuilt the way they think, who are AI-capable in their sleep. Same rollout, wildly different outcomes, all under one roof. Which is why adoption figures flatter everyone. Ireland’s 92 per cent measures whether the tool is in the building. It says nothing about how many people in the building can use it with judgement, and that spread, not the subscription, is where the value gap lives.\n\nIt also explains one finding in the Irish research that deserves more respect than it has been given. When Google and Amárach asked 400 Irish SMEs this spring what held them back, the most common answer was fear of making mistakes, ahead of a lack of skills and well ahead of cost. That gets reported as hesitancy, a nation of cautious adopters trailing the curve. I’d read it as the most clear-eyed finding in the study. Those owners have sensed, correctly, that they don’t yet have the judgement in-house to catch AI’s errors, and they are declining to deploy a tool they can’t supervise. The instinct is sound. What it calls for is building the capability that makes the caution unnecessary, and no subscription will do that on its own.\n\nBe suspicious, too, of anyone selling the alternative. The 100 per cent successful AI rollout, where everything worked, and everyone levelled up together, is a Jackanory story. We are inside a transformation with too many moving pieces for that: people move at different speeds, with different mindsets, at different levels of adoption, and no rollout brings everyone through at once. The factory engineers of the 1920s never found the perfect layout either. They built floors that could be rearranged as the machines and the people kept changing, and the rearranging never stopped. Whatever you redesign this year will be partly wrong by next year. Waiting for it all to settle means waiting forever. The realistic aim is a version you expect to revise, which still beats a bolt-on that fails in a way we’ve understood since the dynamo.\n\nSo before the next purchase, or the next renewal, one question: is this a bolt-on or a redesign? What work will we rebuild around it, and who will we develop to run it? If there’s no answer, you’re not adopting AI. You’re paying for it.\n\n*Rahim Hirji is a future-of-work strategist who helps organisations stay human as technology changes faster than people can adapt. He is the founder of The SuperSkills Intelligence Company and author of SuperSkills: The seven human skills for the age of AI, published by Kogan Page.*\n\nSee more breaking stories [here](https://irishtechnews.ie/category/cutting-edge/).\n\n```\nMore about Irish Tech News\nIrish Tech News are Ireland’s No. 1 Online Tech Publication and often Ireland’s No.1 Tech Podcast too.\nYou can find hundreds of fantastic previous episodes and subscribe using whatever platform you like via our Anchor.fm page here: https://anchor.fm/irish-tech-news\nIf you’d like to be featured in an upcoming Podcast email us at [email protected] now to discuss.\nIrish Tech News have a range of services available to help promote your business. Why not drop us a line at [email protected] now to find out more about how we can help you reach our audience.\nYou can also find and follow us on Twitter, LinkedIn, Facebook, Instagram, TikTok and Snapchat.\n```\n\n", "url": "https://wpnews.pro/news/youre-not-adopting-ai-youre-paying-for-it", "canonical_source": "https://irishtechnews.ie/youre-not-adopting-ai-youre-paying-for-it/", "published_at": "2026-07-22 09:30:00+00:00", "updated_at": "2026-07-22 09:55:51.084843+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-policy"], "entities": ["Trinity College Dublin", "Microsoft", "Rahim Hirji", "Paul David", "Google Cloud", "DORA"], "alternates": {"html": "https://wpnews.pro/news/youre-not-adopting-ai-youre-paying-for-it", "markdown": "https://wpnews.pro/news/youre-not-adopting-ai-youre-paying-for-it.md", "text": "https://wpnews.pro/news/youre-not-adopting-ai-youre-paying-for-it.txt", "jsonld": "https://wpnews.pro/news/youre-not-adopting-ai-youre-paying-for-it.jsonld"}}