Half of our house is without a roof at the moment.
The house is a 1910 build and we’re renovating it while living in it. Electrics done (mostly). Replastering done. Plumbing pending. The roof came off after we found the leaks the survey missed. Officially the house is ‘livable’. In practice, dinner happens huddled around whatever space we can find that week.
Every conversation I’ve had with a finance leader about AI this year sounds exactly like our house.
One workflow rewired and working beautifully. Another stripped back to the brick. A tool bought and abandoned in the corner like a paint shade that looked different on the wall. And running underneath it all, the same familiar question, asked in the same slightly lowered voice…
“Am I behind?”
What the room actually said
Last Tuesday we ran our AI in Finance webinar. Around 120 finance leaders signed up, and we polled the room on where they actually are with AI. Of those who answered:
47% are embedding AI into one or two workflows.
29% are experimenting on their own.
24% are using it across a few finance processes.
Two numbers matter more than any of those… Not one respondent said they hadn’t started, but no-one said they were done either.
One webinar poll isn’t a research institute, so hold the exact figures lightly. But the shape matches every roundtable and every client conversation we’ve had this year, and the shape is the point. The market hasn’t split into leaders and laggards. Almost everyone is standing in the same half-renovated hallway, comparing themselves to a finished house that doesn’t exist.
Why everyone feels behind anyway
Because nobody posts the scaffolding.
Your feed shows finished kitchens. Agents running the close. Ninety-second demos where the board pack writes itself. What it doesn’t show is the six weeks of data cleaning behind the demo, the workflow that fell over in month two, or the very senior person still checking every number the machine produces, as they should.
Meanwhile the honest picture, from our polls and from every CFO room we’ve hosted this year, is far less cinematic. The gains are real but concentrated in reporting and analysis. The end-to-end, no-hands finance function exists mainly in keynotes. The middle is where everyone lives.
So if you feel behind, check what you’re measuring against. A building site always looks bad next to a brochure.
The middle is a hiring problem in disguise
Most job specs haven’t caught up with any of this. The half-built years need a particular kind of person, and hardly anyone is hiring for them on purpose.
Hire for the team you had in 2019 and you get people brilliant at work the machine is steadily absorbing.
Hire for the 2036 org chart, all agent-orchestration and data leads, and you’ve bought the furniture before the roof is on.
The people worth their weight right now sit in between. They can run the close and question how the close gets run. They’ll take a five-minute AI analysis and check it the way a surveyor checks a wall, looking for what’s wrong with it rather than admiring it. They’re comfortable working in a house where some rooms are finished and some have no ceiling, because that’s the actual job now, and will be for years.
That profile is rare, and it doesn’t look spectacular on paper. It looks like three or four years of proper finance grounding plus curiosity plus a certain calm in ambiguity. Miss it in a CV, and you will, because CVs weren’t built to show it.
What this means in practice
One. Sequence like a builder. Electrics before decorating, data before agents. If the foundations aren’t trusted, every clever tool on top just redecorates a wall that’s about to come down.
Two. Pick the next room deliberately. One workflow, done properly, reviewed honestly. The 47% embedding one or two workflows aren’t behind the 24% running several. They may simply be doing it in an order that holds.
Three. Change what ‘where are we with AI’ means at board level. Swap the maturity score for a renovation update: what’s done, what’s live, what’s fenced off for later, and what it means for who you hire next.
Four. Write the middle into your next job spec. Ask candidates what they’d check before trusting an AI-produced analysis, and listen for whether the answer comes from somewhere real.
Where this leaves you
‘Behind’ implies a finish line, and there isn’t one. Houses like ours don’t finish, either. You get more rooms working, and you stop apologising for the cement mixer. Then, by the time you’ve finally ‘finished’, trends change and its time to redecorate.
So maybe retire the question that’s been sitting under every AI conversation this year, and swap it for a better one. Which room next, and who’s holding the torch while we work?
That question has an answer. ‘Am I behind?’ only has anxiety.
If you’d like a second pair of eyes on where your finance team actually is, and what to hire for from halfway, drop me a note at [email protected].
It’s the Finance Team Blueprint conversation we’re having most weeks with CFOs of scaling businesses, and there’s usually more comfort in the honest picture than people expect.
