Your First 90 Days Leading AI Transformation
(Aka...what to do when you get thrown into the role)
Last week I was on a call with someone whose business card still says VP of Operations. Three months ago she was given a new responsibility that didn’t come with a new title, a new headcount, or a new budget. She’d been told to ‘lead AI transformation’ for her division. She had a 90-day deliverable and a CEO checking in weekly.
The People Leading AI Transformation Are Often ‘Voluntold’
About 40% of large enterprises have a Chief AI Officer or are actively hiring one. That number gets quoted a lot. The piece nobody mentions is what’s happening in the other 60%, plus inside the divisions of the 40% that do have a CAIO.
That’s where most of the work is being done. By VPs of Operations, Heads of Innovation, Directors of Transformation, line-of-business leaders who happened to be the AI-curious one in the room when the CEO needed to point at someone. Birju Shah, who ran AI at Uber and now teaches at Kellogg, makes the case bluntly: most Fortune 500s don’t actually need a CAIO. They need to retrain the executives they already have. I think he’s right, and I’d add this:
Most leaders are doing the work of AI transformation without a playbook, a team, a budget line, or the recognition. They are running the transformation while people still argue if this should be a role.
When did your organization last clearly define who owns AI transformation, versus who was politely asked to lead it?
The Trap Is Structural, and the Numbers Tell You So
Here’s what people in this position are walking into. Executives expect AI to deliver something on the order of 32% productivity improvement and 26% faster revenue growth. Actual revenue per employee, across the same enterprises pouring money into AI, is up about 1%. The HFS analysts who published that gap recently argued that if your CAIO is still relevant in three years, something has fundamentally failed.
Whatever you think of that conclusion, the gap itself is the part that matters. A 30-point spread between what leadership expects and what the underlying business is producing is a structural problem that any newly tapped AI leader inherits the moment they say yes.
The Bernard Marr case from last year captures the pattern. A Fortune 500 hired its first CAIO with significant fanfare. Eighteen months later they reposted the role. The stated cause was a $50M cost-savings target inside the first year. The actual cause was that the target was never realistic, but it was the basis on which the role had been justified to the board.
Additional Reading: WHO OWNS AI TRANSFORMATION?
I see this pattern most often in people who don’t have the title. They were given the mandate informally, often as an ‘and also’ on top of their existing P&L. They inherit the same expectations gap as a CAIO would, with less authority, smaller teams, and no protected runway.
If This is You, The First Job Is Renegotiating the Mandate
The leaders I see surviving this role 18 months later, across a wide range of titles, did not start by auditing stakeholders, aligning with strategy, or building a quick-wins portfolio. They did those things. But they did them after a different first move.
They renegotiated the mandate.
In practical terms, that meant going back to the executive who handed them the responsibility and getting three things on paper before doing the work:
A clearly named scope. Which functions, which P&Ls, which decisions are inside this mandate, and which are out. Most informal AI mandates begin as ‘go figure it out,’ and that breadth is precisely what makes them fail.
A success definition that isn’t a dollar figure pulled from a board deck. The Marr case is instructive. If the target is “$50M in savings inside 12 months” and that number was generated by someone other than the people who would deliver it, the right first move is to put a different number on paper.
An authority statement. Not a budget. An explicit acknowledgment of which decisions the AI lead can make without escalating, and which require the named executive’s sponsorship. This is the piece informal mandates almost always skip.
Additional Reading: Your Org Structure Doesn’t Like AI
The leaders who get this kind of clarity set themselves up for success. The ones who skip it, even excellent ones, set themselves up for ambiguity at best, failure at worst.
Final Thoughts
Many people I chat with are tolerant of the role ambiguity because it feels like being on the front line. It feels like a failure now will turn into a future success.
And it may.
But why not set yourself up for success from the get-go? (This applies even more sharply when you don’t have the title, because the role wasn’t designed at all. It accumulated.)
If you’ve been handed an AI transformation in the last six months, formally or informally, what was your situation? Were you handed a title, a wish, or a mandate?
Here’s the Good News
If you think that you need to do this without a playbook, I have good news! My upcoming book Hyperadaptive provides a research-backed blueprint across nine dimensions.
Three ways to start before you decide whether to buy.
Listen to a sample of the audiobook → Audiobook sample
Read the opening chapter as a free excerpt → Read the excerpt
Or, if you’re already in: get the book plus the launch-week Reader Bundle, which includes early access to frameworks, the AI Integration Handbook, the Hyperadaptive Model PDF, and 90 day access to the paid Hyperadaptive Intelligence community.
→ Get the book + Reader Bundle
Hyperadaptive: Rewiring the Enterprise to Become AI-Native releases May 12th.