Case 01 · AI-accelerated delivery · Fortune 100 R&D
From 6 months to 3.
The short version
A Fortune 100 R&D division needed an AI orchestrator defined, with no clear brief. I built interactive AI prototypes and put them in stakeholders' hands in session one. Alignment landed in three sessions instead of six-plus months — and the follow-on stadium-operations platform got the fastest approval the account had seen.
The fog I inherited
Nobody agreed on the problem. Three groups, three definitions of "orchestrator", and a delivery calendar already running.
Day 1 · AI orchestrator engagement · Fortune 100 R&D
Most engagements start planning here. We started by building. →
What we knew
The division wanted an AI orchestrator — in principle.
Alignment historically took 6+ months per initiative.
Multiple stakeholder groups, competing priorities.
AI does the heavy lifting; humans stay in the loop.
What was assumed
Stakeholders could evaluate the concept from decks.
Requirements could be defined before building.
"Alignment" meant sign-off on documentation.
The counter-bet: show the experience in action — a live workflow they could get excited about. Proved right.
Nobody knew
What the orchestrator should do on day one.
Who the real decision-maker was.
What would make stakeholders commit.
What an agent orchestrator even looks like in R&D.
Key decisions
Decision 01
Working prototype over concept deck
Traded polish and coverage for reaction speed.
Decision 02
Demoed unfinished software live
Traded control of the narrative for honest signal.
Decision 03
Three sessions with real users present
Traded stakeholder comfort for speed to truth.
Influence
The resistance was real: prototypes over-promise to the client. I scoped every demo as a question, not a commitment — and said so out loud, first.
Reflection
I would bring engineering into session one next time. The prototypes bought alignment we then had to re-earn at build.