Cadence
1 to 3 days a week
From $12,000 a month for one day a week, 6-month minimum, quarterly renewal after. The day mix follows your board cadence, your exec-staff load, and the first 90-day shipping target.
Home / Hiring an AI leader
For companies hiring an AI leader
The AI seat is still empty.
You posted the role 45 days ago, maybe 120. The right resume has not come in. I can own the AI program starting now while you keep hiring, so the next exec inherits a running system instead of a backlog.
Book 30 minutes with Edwin
Not ready to talk? Score your AI program in about four minutes.
- Start$2,500The plan, two weeks
- ThenFrom $12kA month, Fractional retainer
- Cadence1 to 3 daysA week, 6-month minimum
- Exit14 daysMutual, from day one
01 · The moment
Where thesearch is stuck.
The role has been open 45, 90, sometimes 120 days. Your recruiter is grinding a cold pipeline. The shortlists keep coming back the same way. Senior candidates who pass the bar are already running AI orgs and are not looking. The ones who are looking either lack shipped production credit, or want a title and comp jump the board is not ready to underwrite.
Every week the role stays open, the board asks the same question and gets the same deck. The teams that need AI cover do not get it. The quarter slips and the plan slips with it.
Talent concentration
The people who can do this job are already running AI orgs at the labs and at AI-native companies. They are not opening recruiter InMails, and the usual executive pipeline misses them entirely.
Comp ceiling
Your AI exec budget is anchored to a pre-AI compensation reality. Candidates who pass the technical bar price themselves against AI-native comp bands. The board approved last year's number.
Board pressure
The board wants an AI story this quarter. Every quarter without an exec, you are defending a search when you should be defending a portfolio.
Role drift
The shape of AI leadership has moved twice since the job description was written. Sometimes the candidate gap is really a spec gap.
02 · Why fractional
The gap isnot idle time.
While the seat is empty, your AI work runs without an owner, and that has a shape you can predict.
You announce that the company is going to use AI. Very little of that announcement reaches the floor. Gallup found only 22% of employees say leadership has communicated a clear plan for AI, while 44% say the company has already started integrating it (June 2025). The further someone sits from the decision, the less they know. In a later Gallup survey, 26% of individual contributors could not say whether their organization had implemented AI at all, against 7% of leaders.
People do not wait. They start on their own, on personal accounts. MIT's NANDA project, in its State of AI in Business 2025 report, found workers at roughly 90% of the companies it surveyed using personal AI tools for work, against about 40% of companies holding an official subscription. Whether that work gets shared is a social question. Slack's Workforce Index found 48% of desk workers would be uncomfortable telling their manager they used AI. The top reasons were that it feels like cheating or makes them look less competent, and company policy ranked dead last. A KPMG and University of Melbourne study of 48,000 people across 47 countries found 57% hide their AI use and present the output as their own.
So you get pockets. Some people build something good and never tell anyone. Others build the same thing twice. A few ship something into production that nobody reviewed. Retool surveyed 307 CTOs, CIOs and CISOs in May 2026. 75% are working under an AI directive from leadership, 93% worry about vibe-coded tools reaching production, 19% confirm an AI-generated tool has already caused a production incident, and only 8% call their internal-tool governance strong. Duplication ranked second to last on their list of concerns, which is usually what a problem looks like right before it gets expensive.
The scarce skill is knowing what is worth building, and volume and judgment come apart quickly. METR ran a randomized trial where experienced developers were 19% slower with AI tools while believing they had been 20% faster. UpGuard found executives had the highest rate of regular unsanctioned AI use of any group. Your heaviest AI user is not automatically the person who should decide what your company builds.
That ownership gap is what fractional covers. Someone decides what is worth building, sets the guardrails before people are already in production, and gives the pockets of good work somewhere safe to surface. I work the search alongside your recruiter, rewrite the job description against what we learn, and step out when the right person signs. A role that is not on fire gets a better hire.
Sometimes you do not need a leader at all. A workshop on tools and practices is often enough when the wins live inside one team. Your tenured people learn the tools and carry them, because the floor already trusts them. You need a program owner when the work crosses functions and someone has to tie the initiatives to business outcomes.
The counterweight. Wharton's 2025 adoption study found 74% of enterprises that measure gen AI ROI report positive returns. Outcomes are bimodal, and the split tracks execution discipline far more than model choice. That is the argument for owning the program now instead of waiting for the hire.
03 · The cost of waiting
What a wrong hireactually costs.
Commonly cited estimates put the fully loaded cost of a failed senior tech exec hire at roughly 2 to 3 times annual comp. That covers search, onboarding, the work that did not ship, severance, and the second search. For a Director or VP of AI, that is a high six-figure to seven-figure mistake.
Settle on a hire
- Senior exec comp, fully loaded.
- Six months of ramp before anything ships.
- Risk of an exit around month 14, when the spec mismatch becomes obvious.
- A failed hire costs 2 to 3 times annual comp.
Bridge with GAT →
- $2,500 for the plan. Two weeks, and it doubles as the research for the job spec.
- From $12,000 a month after that. That buys one day a week, up to three days, 6-month minimum, with the $2,500 credited in full.
- The search keeps running. You hire right, or you learn the role has changed.
- A 14-day mutual exit from day one.
These are ranges, not promises. The point is the order of magnitude. Six months of the bridge costs a fraction of one wrong hire, and you keep the search running the whole time.
04 · How it starts
The plan first.Then keep me on.
Every engagement starts with the same two weeks, even when you already know you want the retainer. The plan tells both of us what the job actually is, and that makes it the best input you will get for the job description too.
-
01
The plan: AI Program Diagnostic
A workflow map built from your system data and interviews with the people who do the work. A 90-day sequence with a named owner, a measured baseline, and a build type on every line. A kill list. A 60-minute readout on day 10.
Flat fee, paid before kickoff. If the readout slips past day 10 or any line is missing an owner, a baseline, or a build type, you get the $2,500 back. Assumes you attend the kickoff and give me the access on the kickoff list.
$2,5002 weeks -
02
Keep me on: Fractional Head of AI Transformation
I own the program and the next builds. I sit in the exec staff, take the AI portfolio to the board, and work the search with your recruiter.
1 to 3 days a week, 6-month minimum, quarterly renewal after. 14-day mutual exit from day one, 30 days notice after month three, days not worked refunded.
From $12kA month
Full creditStart the Fractional retainer within 30 days of the readout and the whole $2,500 comes off it. If the plan says one workflow should ship first, the Scoped AI Build takes the same credit.
05 · What you get
What the next execinherits.
Everything here outlasts me and belongs to whoever takes the seat.
- An AI portfolio the board can see. Every bet classified Now, Next, or Later, with the reasoning written down. "What is our AI plan" gets a real answer.
- One production AI use case shipped in the first 90 days. In production with an eval bar, used by the team, measured against the baseline in the plan.
- A redesigned role description. What the job description should say, based on what we learned. Sometimes the right next hire is two ICs and a contractor, not one Director.
- Vendor and tooling decisions, documented. The next exec inherits a vendor stack, contract terms, eval data, and the reasoning behind each choice.
- An eval bar the org already works to. Eval suites, prompt regression tests, and safety reviews, so the team knows what good looks like before the new exec walks in.
- A team that uses AI in the open. Workforce training in three layers. Leadership signals that AI use is safe to admit, a small lab group runs the experiments, and everyone else learns the few moves that matter for their job.
- Guardrails people can follow. What AI can decide alone, what needs a human on it, which data never goes near a model, and where the sanctioned tools live. Written for the person doing the work, so unsanctioned use turns into visible use.
06 · The engagement
Built asa bridge.
Exit
14 days, then 30 days notice
A 14-day mutual exit from day one. After month three either side can exit with 30 days notice. Days not worked are refunded, so the bridge ends clean.
Handoff
Built in from day one
When the full-time exec signs, fractional ends or steps down to an advisory role. The portfolio, the decisions, and the standards are already written down.
Search partnership
I work with your recruiter
Job description review, panel interview support, evaluation rubrics, and reference calls. What I learn in the seat shapes who we hire into it.
Capacity: I run three diagnostics at a time, and builds plus retainers are capped at four client commitments. The next diagnostic opening is October 6.
07 · FAQ
Questions.
Does fractional compete with our full-time AI exec search?
No. I work the search with you, help write the job description against what we learn together, and hand off cleanly when the right exec signs. Fractional engagements like this usually end inside 9 months, when the full-time hire lands.
Why start with the $2,500 plan instead of the retainer?
Because the plan tells both of us what the job actually is. In two weeks you get a workflow map, a 90-day sequence with an owner, a baseline, and a build type on every line, and a kill list. That is also the best input you will get for the job description. Start the Fractional retainer within 30 days of the readout and the whole $2,500 comes off it.
What if we hire someone fast?
Then we ramp down. The Fractional retainer carries a 14-day mutual exit from day one. After month 3 either party can exit with 30 days notice. Days not worked are refunded. The goal is to make your AI exec successful from week one, and the bridge ends when they are ready.
How is this different from an AI strategy consultant who would help us hire?
A strategy consultant writes the deck and leaves the seat empty. I sit in the exec staff, own the AI portfolio in front of the board, ship one production AI use case in the first 90 days, and hand the keys to the full-time hire. There is a side-by-side comparison if you want it.
What if the role we're hiring for is wrong?
That happens. An AI role spec written in 2024 is often the wrong spec for 2026. One artifact of the engagement is a redesigned role description based on what we learn together. Sometimes the right next hire is two ICs and a contractor, not one Director.
Do we need to hire an AI leader, or is training enough?
Sometimes you do not need a leader at all. A workshop on tools and practices is often enough when the wins live inside one team. You need a program owner when the work crosses functions and someone has to connect the initiatives to business outcomes. The Diagnostic tells you which situation you are in.
GAT.
Keep hiring. Stop losing quarters.
Tell me where the search stands. I will tell you whether the bridge fits, and what the first two weeks would cover.
Book 30 minutes with EdwinNot ready to talk? Score your AI program first. Twelve questions, about four minutes, scored breakdown by email.