AI is starting to show up in the growth numbers, but inside most organisations, we are still measuring the wrong thing.
Last week, two BBC articles dropped that perfectly encapsulate why we built AcademyAI.
First, the macro. The ONS said the UK economy grew 0.4% in July, against forecasts of no growth at all. Services led business and computer programming were the standouts.
Liz McKeown, the ONS director of economic statistics, said there was evidence that businesses involved with AI and related technologies helped to boost the sector across May, June and July. In addition, many of the IT firms reporting the largest turnover, the ONS noted, "appear to be involved with AI", even if the exact contribution is hard to pin down.
So the national accounts are beginning to notice AI, even if they cannot yet put a clean number on it.
Promotion is already tying to usage
The second article in the same week showed how AI was showing up at an individual employee level. A separate BBC piece asked whether promotion should depend on how workers use AI. The answer, in a growing number of firms, is already yes.
Accenture's Julie Sweet put it bluntly: "Today, AI at Accenture is how we do work. So if you want to get promoted, you've got to do the things that we do in order to operate at Accenture".
Disney, Meta, JP Morgan and KPMG have been reported as running AI leaderboards that rank employees on usage. Coinbase has already fired engineers who did not complete AI training. Employees describe a quiet two-tier workforce: same title, same tenure, different value depending on whether someone treats AI as a threat or a tool.
On the surface, that sounds like progress. Leadership is serious, incentives are aligned, and usage is rising. But if you look closer, you can see the same plateau we keep writing about here.
You will measure adoption, not judgement
Kamila Miller, applied AI researcher and lecturer at Henley Business School, named it perfectly in the BBC article: make AI usage a KPI and people will log interactions to hit the metric, route work through a chatbot that did not need it, and generate AI-flavoured outputs that look productive on a dashboard. "You will measure adoption. You will not measure judgement, learning, or better decisions". Her closer: "You have not made people more skilled - you have made them more obedient"
Duolingo and Amazon have already walked some of this back after people started using AI for AI's sake. The usage metric was never the goal. It was a proxy for a thing we have not defined.
Will does not equal skill
Which brings me back to last week's Live with Erica Farmer, and to the reason AcademyAI exists.
Erica's point was not that people should ignore AI. It was that tool access does not make people AI literate, that "will does not equal skill", and that adoption is not capability. She would rather have 20% of an organisation ideating with AI as a thinking partner than 100% rewriting emails with Copilot. Most organisations, we said on the Live, are still stuck in personal productivity: the best people get faster, the gap widens, and the wins of the few rarely become organisation-wide efficiency, let alone the harder work of redesigning how work gets done.
We are only at the very earliest stages of understanding what "good with AI" looks like, and what using AI well actually means at work.
Right now we are surrounded by anecdotes and some usage data. Licence counts, training completions, tokens and leaderboards. Along with anecdotes of the exciting agent someone in another team built.
What we do not have, in most organisations, is an objective measure of capability.
That is the most important reason we created AcademyAI.
You cannot measure what you have not defined
Understanding AI is becoming as fundamental as understanding how to use a computer at work. It will shape how businesses grow. It will shape how people grow their careers. It will shape how organisations contribute to the economy. If the ONS is already picking up AI's impact on the economy, and if employers are already tying promotion to AI use, then guessing at capability is no longer a soft HR problem. It is a growth problem, a fairness problem, and a leadership problem.
You cannot manage what you will not measure. And you cannot measure what you have not defined.
Our work is to make workforce AI capability visible: assess the gap, close it by role, and prove it moved. A clear picture of the skills that matter when intelligence is abundant, and a way to lift the baseline so the gains stop living in private pockets.
To close out with one of my favourites from my chat with Erica: "If you are not sharing so the organisation can benefit from the great work you are doing individually, that is as bad as not doing it at all"
The economy is starting to score AI. The question for every leadership team is whether they are scoring the thing that actually creates value, or just the theatre of usage.
If this landed, reply and tell me where your organisation is still guessing.
Sources
- Emer Moreau, "AI boom helps drive surprise UK growth in July," BBC News, 11 Sep 2026, https://www.bbc.co.uk/news/articles/cq5xjlvn71lo
- MaryLou Costa, "Should promotion depend on how workers use AI?," BBC News, 9 Sep 2026, https://www.bbc.co.uk/news/articles/c1j1896e973o
- AcademyAI LinkedIn Live with Erica Farmer, 10 Sep 2026, https://www.linkedin.com/events/7500451560234147841
