Twenty years of finance transformation, and the strategy document has barely changed.
That was one of the first things Anders Liu-Lindberg put to me on our upcoming episode of The Conscious Finance Podcast, and it’s been rattling around my head ever since. Put a finance strategy from today next to one from the mid-2000s, he reckons, and they look almost identical. Except now it says something about AI.
Which stings a little, because he’s right. Finance has described itself as ‘in transformation’ more or less continuously since the big ERP systems went in. And yet, in his words, the business doesn’t care much more about finance than it did twenty years ago. Most of the effort has pointed inwards. Faster closes, better processes, improvements for finance, presented with a straight face as improvements for the business.
Underneath every AI plan I’ve seen this year sits the same unasked question. Is this for ten percent faster, or for ten times better?
Two very different ambitions
The distinction belongs to Jeremy Utley, the Stanford adjunct who’s spent the last couple of years urging leaders to aim for ten times better with AI when most settle for ten percent. Anders and I borrowed it for a good chunk of the episode, because the forecast makes a tidy example.
Ten percent better looks like the same forecast, delivered a day earlier, with fewer people involved. He’s careful not to sneer at that, and so am I. It’s good. Time back is time back, and if your team is drowning, time back is everything for a while.
Ten times better is a different animal. A forecast that runs a thousand scenarios and shows you where you’re most likely to land, instead of one number and a confident nod. One that’s plugged into external data as well as your own, sits closer to real time, and lets you play out scenarios in the meeting rather than three weeks after it. Capabilities you don’t currently have, rather than a faster version of the ones you do.
Same words on the project plan, but a completely different destination.
What ten percent thinking costs
At the start of August, PwC published a survey of just over a thousand senior executives at US financial services firms. 77% said most of their AI investments are not delivering measurable ROI. Nearly eight in ten expect their workforce to shrink by at least 20% within five years. And of the firms doing proper modelling of AI’s impact on their people, only half have even looked at redesigning the work itself.
If you read those together, a picture starts to form… Firms buying speed, cutting seats, and struggling to point at the value. The car is lighter, but it’s still the same car.
A slightly cheaper version of the same finance function was never going to show up as transformation on anyone’s dashboard. Ten percent is a saving you bank once. Ten times compounds.
Why almost everyone is stuck at ten percent
Anders had hoped 2026 would be the year finance moved from experimentation to scaling, and from what he’s seeing across clients, it isn’t happening yet. The governance isn’t in place, cybersecurity keeps the brakes on, the data foundations aren’t there, and the business case is fuzzy. So teams build agents on top of their current stack, automate a bit of forecasting, and collect small wins scattered all over the place.
None of that is wasted. But it’s improvement, and improvement was never going to reimagine anything, in the same way a service and an MOT will keep your car running beautifully without ever turning it into something else.
Then he mentioned a request that landed with his team recently… A large company asked, in effect: if you were designing finance from scratch in the world of AI, no legacy, what would it look like? Anders called that the basis for true transformation, and I think he’s right.
The two camps
From my side of the table, CFOs are settling into two camps…
Camp A is racing for the ten percent. Headcount and hours saved. Understandable, especially with investors asking pointed questions about cost.
Camp B is asking a different question altogether: how does this create value? How does it get us to our plan, or our exit, faster? How does it make the business worth more?
Both camps are using the same tools. And in fairness to the first, ten percent is the easier case to write down, which is exactly why it keeps winning… and, if PwC’s numbers are anything to go by, exactly why so few can point at the return afterwards.
The team consequence
There’s a hiring decision buried in all this, which is why a recruiter is writing about AI strategy at all.
The ten percent road keeps the old shape of the team, slimmer. It usually ends in the same hire you made last time, minus one.
Ten times changes the shape. A forecast that runs a thousand scenarios produces a thousand answers, and someone has to know which assumptions are creaking and how to walk a room of non-finance people through the three that count. The machine multiplies the output. Judgement, translation and trust turn the output into a decision, and on the ten times road that work fills the whole job description.
I’ve watched a team trade two transactional seats for one finance business partner and end up with more influence than they had on the bigger headcount. The board asks them different questions now. The saving paid for it, and it could just as easily have become a cost reduction nobody remembers.
I’ve talked clients out of hiring before now, which is mental, for a recruiter. This is usually where it happens: the like-for-like replacement of a seat the ten times version would have redesigned.
What this means in practice
One. Keep the ten percent projects running. They fund the journey and buy credibility for the bigger ask.
Two. Pick one process this quarter and ask the blank-sheet question of it properly. The forecast is the strongest candidate: what would this look like if we designed it today, for this business, with these tools?
Three. Decide now what a saving becomes when it lands. A cost reduction, or the budget for the business partner you could never justify before. Write it down before the saving arrives, because afterwards the default answer is always ‘cost’.
Four. Read your next finance job spec and ask which version of the function it describes. A spec written for the ten percent world will staff you for it. If it’s a like-for-like replacement of a seat the ten times version would redesign, give it a month and do the thinking first.
Where this leaves you
You’ll probably never build the from-scratch version. Legacy is real, audits are real, and nobody hands you a greenfield finance function on a Tuesday. But the blank-sheet question is still worth an hour of your time, because the gap between your answer and what you run today is your actual transformation agenda. Everything else is maintenance.
Ten percent faster and ten times better start with the same software and the same steering group. The difference is the ambition in the room, and ambition, unlike software, doesn’t come with an implementation partner. It comes with the people you hire and what you ask of them. The redesigned car needs a different kind of driver.
So it might be worth putting the question that company asked Anders to your own team. If we were building finance from scratch, what would we build? The answers tend to be braver than the roadmap.
The full episode goes deeper, into forecast credibility, diamond-shaped teams and where the next generation learns. Coming soon… watch this space!
And if you’re staring at this fork with a live decision in your hands, drop me a note at [email protected]. Happy to share what we’re seeing work, and where the wrong turns are. The conversations that start with a blank sheet are usually the best ones.
