Sort it out internally
It lands on whoever has the least on their plate that month. They can map their own function and none of the others. The dependencies between teams are exactly the part nobody can see from inside one team.
AI Program Diagnostic · Two weeks
You have AI running in several places and nobody owns the program. In two weeks you get the map and the sequence.
Pilots, vendors, a few integrations, and no single person accountable for how they fit together. The Diagnostic maps every owner, integration point and vendor risk, then sequences the next 90 days with a name against each line.
TL;DR · 60-second read
What you walk away with
A dependency map across everything AI you have running, and a 90-day plan sequenced so the pieces stop blocking each other. Two weeks.
- Who it's for
- You own the AI decision. You have tried at least one use case. Nobody is running the program. Usually 20 to 200 people.
- Format
- Two weeks. Kickoff call, the map and the sequence, readout call.
- Investment
- $2,500, flat. Paid before kickoff.
- Next step
- Book the Diagnostic
The problem
The pilots worked. The program stopped.
None of this is a model problem. Two teams are waiting on the same integration. A vendor renewal is coming and nobody has read the contract. The thing that shipped in March has no owner. Every piece has a champion. The program has nobody.
Hire a Head of AI
Research this year puts a full-time AI leader out of reach until roughly $100M to $150M in revenue, and most sources say under $300M cannot justify one. The search runs two quarters. The program sits still for both of them.
The fix
Rent the owner for two weeks
I map what depends on what, then sequence the next 90 days against it with a name on every line. You get the artifact and the owner list. Not a study.
What you get
Three things in two weeks. A call to start, the artifact, a call to land it.
Day 1 · 60 minutes
Kickoff call
A working session. I pull the inventory out of your head: what is running, who touched it last, which vendors are in play, and what your CEO is actually asking for.
Days 2 to 9
The map and the sequence
I trace the dependencies across owners, integrations and vendors, then build the 90-day sequence against them. I come back with questions when the map does not close.
Day 10 · 60 minutes
Readout call
We walk the map. I answer your questions and rehearse the pushback so you can defend the sequence in front of your CEO and whoever holds the budget.
The artifact
Five parts. Every one names a person or kills a line item.
No filler. Each part answers a question you will get asked before you get the chance to ask it.
Part 1 · The core
Dependency map
Everything AI-related you have running, on one page, with owners, integration points, vendor risk and downstream impact. This is the view nobody inside a single function can produce.
Part 2 · The blockage
Where the program is stuck
The places two workstreams are waiting on each other, and the ones waiting on a decision nobody has been asked to make. Named, with what it costs to keep waiting.
Part 3 · The differentiator
Where AI is a trap
The two or three moves to kill. The chatbot every vendor is pitching, when the money is in back-office exception handling. Most plans only say what to start. This one says what to stop.
Part 4 · The path
90-day sequence, with names
What happens in the next 30, 60 and 90 days, ordered so the pieces stop blocking each other. Every line has a human accountable for it. Sorted against a three-horizon frame so you can see what is now and what is later.
Part 5 · The next move
What it takes to run it
Whether your team can carry the sequence, what you would have to hire, and where I would be useful after this. If the answer is that you do not need me again, that is in there.
Proof
Getting owners to agree on a trade-off is the job I did at eBay.
I was a 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 across Core Product and Core Technology, eBay's two largest engineering organizations. I did not run those programs. I worked the layer underneath them: mapping what depended on what, surfacing risk before it landed, and getting people who disagreed to settle on an answer. That is the layer your AI program is missing.
The guarantee
The Rework Guarantee.
Your real risk here isn't the $2,500. It's spending two weeks and ending up with a document nobody can act on. So the guarantee is built around the artifact, not around how anyone reacts to it.
The guarantee
Usable or I rework it
If the map and the sequence aren't something you can act on, I rework them until they are. If they still miss, you get 100% back and you keep them.
How it works
30 days, one email
You have 30 days from delivery. Reply to my email and the refund hits your card in 3 business days. No forms. Whether your team runs the sequence, you hire for it, or you bring in someone else, this has nothing to do with whether you work with me again.
Who it's for
Built for one situation. Honest about who it's not for.
This is for you if
You own the decision and nobody owns the program
You can say yes to this without a committee. You have at least one AI use case that already shipped or already stalled. And there is no single person accountable for how the pieces fit. Companies where this lands are usually 20 to 200 people.
You haven't started yet
Nothing is running and nothing has been tried, so there are no dependencies to map. Take the free AI Program Scorecard and come back when something is live.
You're past this stage
You have a Head of AI in seat and the program is already moving. What you need is capacity or a second opinion on a specific bet, not a map. Start a Fractional conversation instead.
Book it
Book the Diagnostic.
I run a maximum of 3 Diagnostics at once. When the calendar is full, booking closes until one wraps.
FAQ
Frequently asked questions
What is the AI Program Diagnostic?
A two-week engagement that answers who owns what in your AI program. You get three things: a 60-minute kickoff call, the artifact, and a 60-minute readout call. The artifact is a dependency map across everything AI-related you have running, covering owners, integration points, vendor risk and downstream impact, plus the places the program is blocked, the moves worth killing, and a 90-day sequence with a named human accountable for each line.
How much does it cost?
$2,500, flat, paid before kickoff. No scoping back-and-forth and no change orders.
Who is it for?
Three things have to be true. You own the AI decision and can say yes without a committee. You have tried at least one AI use case, so it either shipped or stalled. And nobody is currently accountable for how the pieces fit together. Companies where this lands are usually 20 to 200 people. There is no revenue band, because the situation is what matters.
What is the guarantee?
The Rework Guarantee. If the map and the sequence are not something you can act on, I rework them until they are. If they still miss, 100% refund and you keep them. You have 30 days from delivery: reply to my email and the refund hits your card in 3 business days, no forms.
How is this different from the free AI Program Scorecard?
The scorecard is twelve questions you answer about yourself, and it tells you roughly where you stand across ownership, sequencing, accountability and evidence. The Diagnostic is me going and looking. The scorecard scores the situation. The Diagnostic maps it and names who moves first.
Do you build anything during the two weeks?
No. The Diagnostic tells you what to build, in what order, and who owns it. Building happens inside a Scoped AI Build or a Fractional Head of AI Transformation engagement. You can apply 50% of the Diagnostic price as credit toward either if you move within 30 days.
Name an owner
Two weeks from now you know what depends on what, and who moves first.
$2,500. Usable or your money back.