Three stalled pilots and a board deadline, turned into a 90-day plan in four weeks.
A B2B industrial distributor with a new COO and board pressure turns three stalled AI pilots into a quantified 90-day execution plan.
Composite · IllustrativeComposite engagement. The operational shape is drawn from real patterns; the company is not a real named client and the numbers are illustrative. For work that is real and checkable, see baxie.
This composite is a $18M B2B industrial distributor with about 85 employees. A new COO had inherited three in-flight AI pilots, none clearly working, and a board that wanted an AI story before the next quarterly review. The ops team had become the company's integration layer by default, and most of the AI and automation spend was unused or duplicative.
- 01 Four ops people spend about 30 hours a week manually re-entering supplier RFQ quotes, across 35 different PDF and spreadsheet templates, into the ERP. A 6% error rate drives costly misquoted reorders.
- 02 Accounts payable exception routing follows 23 repeatable decision patterns no one had ever written down, eating about 9 hours a week of the AP lead's time.
- 03 The ops team is the integration layer. When the prior head of ops left, three undocumented vendor workflows broke for two weeks and the knowledge walked out the door.
- 04 Tool sprawl across 14 paid AI and automation tools, most of it unused or duplicative.
- 05 Three AI pilots in flight, none clearly working, with a board that wants an AI story before the next quarterly review.
Map the decisions, not the org chart.
Built a decision-flow map across 8 stakeholders, a quantified waste table, an AI-eligible vs AI-ineligible workflow inventory, a 90-day execution plan, and a board-ready one-pager.
Find the two workflows that matter most.
Pinpointed RFQ (request for quote) data entry and accounts payable exception routing as the two workflows with the most to gain, then separated what was AI-eligible from what only looked like it.
Name what lives only in people's heads.
Identified the undocumented vendor workflows living only in people's heads as a continuity risk, and scoped them for capture in the Build phase.
A roadmap the COO can defend.
Delivered a prioritized roadmap of which agents ship first, what stays human, and how to consolidate the redundant tool stack, framed at board altitude.
I'd inherited three pilots and a board deadline. In four weeks I had a map of where the waste actually was and a plan I could put in front of the board, not another pilot.
Three stalled pilots is a diagnosis problem, not a tooling problem. Start with the map.
If you have AI work in flight and no clear read on what is actually working, the audit is the first move. Four weeks to a plan you can defend.
What buyers ask about this one.
Our AI pilot works in a demo but we cannot get it into production. What is actually blocking it?
Almost always a dependency nobody wrote down: a data access approval, a system of record that will not expose an API, or a decision that needs an executive who has not been asked. The diagnostic maps those before it recommends anything, because a pilot that stalls twice usually stalls on the same unmapped thing.
How do you turn stalled AI pilots into a 90-day plan?
Inventory what is running, name the owner of each piece, map what each one waits on, then sequence so nothing blocks anything else. Kill the pilots that cannot clear their dependency inside the quarter. What survives gets a named accountable human per line and a date.
How long does an AI Program Diagnostic take for a 100-person company?
Two weeks from kickoff to the delivered dependency map and 90-day plan. Interviews take the first week, the map and sequencing the second. Larger organizations do not take longer, they take more interviews inside the same two weeks.
How do you evaluate AI vendors when nobody internally can tell the difference?
By writing the evaluation criteria before looking at any vendor, tied to the workflow you are actually trying to change. Then a bake-off on your own data, with a defined pass bar. The failure mode is letting the demo set the criteria.
What does the diagnostic actually deliver?
A dependency map covering owners, integration points, vendor risk and downstream impact, plus a 90-day plan sequenced against those dependencies with a named human accountable for each line. Both are yours, in a format your team can keep updating after I leave.