I was a Technical Program Manager at eBay, reporting to a VP of Engineering and working with directors across Core Product and Core Technology, its two largest engineering organizations. My job was surfacing risk early and getting roughly a dozen teams aligned on outcomes. I ship production AI today as the founder of baxie. Your AI work gets an owner, not another advisor.
You have tried a few things. Something demoed well. Very little of it is doing real work.
AI programs fail the way engineering programs fail, and the failure has nothing to do with the models. It is the dependencies between the pieces that nobody is tracking:
That is a program problem. It gets solved by someone who has run programs, not by buying another tool.
Every engagement starts with the diagnostic. I will not sell you a build before I know what it depends on, and I will not take a retainer from a company that has not tried one AI use case yet.
A dependency map across everything AI-related you have running: owners, integration points, vendor risk, downstream impact. Plus a 90-day plan sequenced so the pieces stop blocking each other, with a named human accountable for each line. The same discipline I applied on engineering programs at eBay, pointed at your AI portfolio. Read the engagement.
The first system from the plan, shipped into the stack your team already runs. No migrations. Guardrails, eval suites, and observability come with it, so your CFO can see what it does and what it costs. Includes the shared knowledge base your team and your AI both read from.
Training so your people run the systems, read the evals, and spot the next opportunity without me. The engagement is not done when the code ships. It is done when your team can take it. Retainer available after handoff if you want the next build.
The terms, before you ask. Diagnostic runs two weeks. Retainers run 1 to 3 days per week on a 6-month minimum, quoted in writing within 48 hours of the call. Every engagement carries a 14-day mutual exit from day one, and after month three either side can leave with 30 days notice. Days not worked are refunded. I hold about four active client commitments at a time, so the calendar is the real constraint.
Technical Program Manager at eBay, reporting to a VP of Engineering. My peers were directors and engineering managers. My day job was getting them to agree on trade-offs across large engineering programs, roughly a dozen teams spanning Core Product and Core Technology, eBay's two largest engineering organizations.
That was the job before AI was production-ready. Owners, integration points, risk, downstream impact. It is the same map an AI portfolio needs, and it is the layer almost every AI consultant skips, because most of them have never run a program.
A senior AI leader is hard to find, slow to hire, and expensive to keep. Most candidates who clear the bar have run a pilot, not owned a program that shipped across a business.
What "fractional" means here: I do the work a full-time AI leader would do, at the fraction of time the work actually needs. Setting direction, aligning the people who have to agree, owning delivery. Without the seat, the equity, or the 9-week search.
An agency hands you a tool and leaves the program unowned. A hire takes nine weeks to find and six months to ramp. I own the program from week one, ship the systems, train your team, and hand it off.
I build. The plan exists so the build lands. You end with working systems and a team that can run them, not a report.
The diagnostic delivers in two weeks. First system in production inside 60 to 90 days from the scope session.
Typically 20 to 200 people. The size matters less than the situation: you own the AI decision, you have tried at least one use case, and nobody is running the program. I have taken that work in professional services, distribution, healthcare software, and single-family offices. Program discipline travels across industries. That is the whole point of it.
On your stack, under your controls. The shared knowledge base is the layer your team and your AI both read from. It is not a place your data leaves, and it is not a product I resell you.
The AI Program Diagnostic is a fixed fee, quoted on the call. Builds are priced to the scope, retainers to the day count, both in writing within 48 hours. I do not bill hourly and I do not take vendor referral fees, so the number does not move once it is written.
Thirty minutes. If there is a fit, I will scope it. If there is not, you will leave with a sharper read on your next move, which is worth the half hour on its own.
Not ready to talk? Score your AI program first. Twelve questions, about four minutes, and you get the scored breakdown by email.