Interim CTO, or Chief Data & AI Officer

Which title applies to you

Two names, one scope, two company stages.

Interim CTO if your product is the data or AI system. In that company the context layer, the retrieval architecture and the agents are not a support function, they are what you sell, and owning them is the technical core of the business.

Chief Data and AI Officer if you already have the scale, the team and the cloud bill, and what is missing is an operating model. Same work. The title changes because your org chart already has a CTO in it and that person owns something else.

What I do not run, in either case: general platform engineering, infrastructure, security or SRE. “CTO” normally implies all four and mine does not. If your hard problem is Kubernetes, incident response, or hiring twenty backend engineers, I am the wrong call and I would rather you knew now.

The engagement exists because a company between roughly 300 and 3,000 people usually needs executive-level ownership of this layer before it can attract a permanent executive, and because the gap left by a departing data leader does more damage in six months than most boards expect.

When companies call me

The seat is empty and a proper search takes six months. The board asked for an AI answer and nobody credible owns it. The cloud data bill became a line item your CFO asks about by name. An audit or an EU AI Act deadline is coming and there is no defensible written position. Or you have twenty people in data, three versions of revenue, and leadership has quietly gone back to spreadsheets.

How I run the mandate

I own it. Budget, roadmap, hiring, vendors, and the story your board hears. An interim who advises rather than decides is a consultant with a longer invoice.

I hire my replacement. The search starts in month two, not month ten. The last month of every mandate is spent with the successor in the seat while I am still reachable, because a handover document is not a handover.

I switch things off, loudly. Retiring unused reporting quietly avoids an argument and teaches nobody anything. A monthly note saying what we turned off and why does more for the culture than any training programme.

What I have actually done in this shape

Four times as an employee, which is a different thing from doing it as an outsider and I say so on the case studies. At HelloFresh I built the BI function in Australia, then the global data organisation, then a data management department. At Babbel I built the first data product organisation and got customer lifetime value and acquisition cost into the hands of the teams spending the money.

The part most relevant to the CTO framing is the most recent. Babbel had no AI agents in production. I shipped the first one as an internal application, then the customer-facing agents that followed it, with engineers given the tooling first and Legal in the room from the first design review rather than as an approval gate at the end. Time from question to answer went from two to four weeks down to five to ten minutes over the same period, because the layer underneath the agents got rebuilt at the same time.

The relevant caveat: each of those came with a permanent contract, a full-time week, and years of tenure. Two days a week for six months is a narrower instrument. What transfers is the judgement about sequence, which is the part that is hard to hire for.

What this costs

Quoted per engagement, based on days per week and duration. I publish a minimum rather than a day rate, because a day rate invites comparison against agency-supplied contractors and the comparison is meaningless: you are buying a decision-maker, not a pair of hands. Below the minimum the scope is too narrow to change anything, and I would rather say that up front.

I do not work through intermediaries on a commission basis.