Most companies cannot honestly prove AI’s value by December, because they never built anything to measure against. Measure what you can today, then use the fourth quarter to build the baseline, so that on 1 January you have real numbers for 2027 instead of invented ones.
Why can’t most companies prove AI’s value by December?
Because they told their people to start using AI and never decided what they would measure. In that situation the only honest number is what you spent. The deadline pressure is real: in a Harris Poll survey of 685 CIOs for Dataiku, published on 24 September 2026, 72 percent could not consistently confirm whether their agents deliver the business outcomes they were built for, and 72 percent said their AI budget is likely to be cut or frozen if performance targets are not met by the end of 2026.
Think about how you would judge any other system change. Nobody judges an SAP rollout in its first quarter. Every employee has to adjust to the new way of working first, and the value shows up after that. AI is a larger change than SAP, not a smaller one.
What can you honestly measure right now?
More than nothing, and less than a return on investment. Almost every company can report three things today: what it has spent, how many people actively use the tools, and three to five real, documented cases. A documented case has an outcome someone can check: a cost saved for a client, a new project won, a license fee avoided because your people rebuilt the software themselves. In this first generation of AI, the return shows up in your people before it shows up in the spreadsheet, so the cases matter more than the totals.
What does an honest AI baseline have to capture?
Three things: how your people actually use AI, what they need to use it better, and which uses produce an outcome someone can check. Three months of that is more than most of your competitors will have in January, and it is the only foundation on which a 2027 AI number means anything.
Why do most AI baselines start with dishonest data?
Because in most companies people hide their AI use, afraid they will get in trouble for it. Ask them what they use before that fear is gone, and your baseline measures what people thought was safe to say, not what they actually do. Removing the fear comes before measuring anything. The goal is a company where people stop hiding their AI use and start being proud of it.
Three months spent inventing numbers leaves you with nothing in January. Three months spent building the baseline gives you 2027. And if the budget freeze comes anyway, remember that cutting before you know where the money went is the wrong lesson.
If you are leading your organization through this, this is the kind of work I do with a limited number of senior leaders each quarter. → https://anglero.com/the-reset/
Strategic AI Advisor, MerkabaPhi AS, Oslo. 450+ keynotes across 30+ countries.
Listen to the Clarity at the Top podcast on Spotify.
Inquiries: anglero.com