Why we fund research that never mentions us
By Polly Barnfield, OBE, CEO of Maybe*
Two years ago we started asking business leaders a question we could not answer ourselves: what does AI actually do in your business?
Not what you have bought. Not what you are planning. What it does, on a Tuesday, to work that matters.
Over 1,500 in-depth interviews later, The Big AI Secret is a research programme with its own name, its own site, and a rule we wrote at the start and have not broken: it never sells anything.
Why a software company runs a research programme
The honest answer is that we needed the research to exist and nobody had done it.
There is no shortage of writing about AI adoption. Almost all of it is vendor material, survey panels, or analyst reports built from the same dozen interviews everyone else used. What there was not, in 2024, was anyone sitting down with a thousand business leaders and asking them what actually happened when they tried.
We wanted to build for the real problem rather than the described one. That meant finding out what the real problem was, and the only way to do that was to ask enough people that the pattern stopped being anecdote.
So we did, and then it turned into something bigger than a market study.
What we found
Adoption was never the hard part. AI has been taken up faster than any business technology in living memory, without a procurement cycle or an implementation project. Almost everyone crossed that line at the same time.
“Fewer than one in five businesses have got past pockets of use to work completing on its own.”
That gap, between adopting AI and integrating it, is the whole finding. And it is not a technology gap. The businesses either side of it are running much the same tools.
It is a gap about whether work starts itself, whether anyone counted what it gave back, and whether the rules are clear enough that people will let AI near something that matters.
The rule we have not broken
The research never sells. Not softly, not in a footer, not at the end of an interview.
There is no pitch on a research call. There is no product in the Guides. Every finding is published free and ungated, with the method and its limitations alongside it. An Analyst who mentioned our software in an interview would be breaking a published rule rather than a courtesy.
People ask why, and there are two reasons.
The first is practical. You cannot get an honest answer to "what has failed already?" from someone who thinks you are about to sell them something. The moment the research becomes a sales channel, the data stops being worth having, and the data was the point.
The second is simpler. The 1,500 people who took part gave us thirty minutes each on the understanding that it was research. Turning that into a pipeline afterwards would be a breach of what we said we were doing.
We disclose that Maybe* funds and runs it, on the site, in the About section and in the footer. It is not independent and we do not claim it is. What it is instead is transparent, which is a harder claim to make and an easier one to defend.
What happens now
The research has stopped being a snapshot. Edition 1 told us where 1,500 businesses stood. It could not tell us whether any of them moved.
So the four stages are now a measurement anyone can take. Six questions, two minutes, and you get an Integration Index and a stage, with a comparison against businesses like yours as more of them answer. It stays on a private page that updates as things change.
The founding cohort is being interviewed again. Nobody has measured the same businesses twice on AI integration, and that is the gap Edition 2 exists to close.
The interviews are conducted by AI Analysts: graduates trained under the published method and credited by name in the research they contribute to. Most of them had never interviewed a chief executive before. They now do it several times a week.
If you want to be in it
Take the two minutes and find your stage, or sit the interview and get an evidenced score rather than a self-reported one. Both are free and neither has anything attached.
And if you can see a hole in the method, tell us. It is published in full for exactly that reason.
Maybe* is the AI execution layer where real business work gets done, with clear rules, full visibility, and your people in control. The Big AI Secret is its research programme, and it does not promote that product.
