Firewall Greg AI prototyping team — company makes no plans against their output

Situation

Peter explicitly framed Gregs AI prototyping work as a separate research division that the rest of the company cannot bet on. No items show on the value drivers board for this work; nothing is committed to customers; deliverables are not on engineering plans. If they emerge with a usable nugget, fine — Wolfords or Nathans team will productize it. Until then, planning treats the team as if it does not exist.

Reasoning

Greg is genuinely valuable as an exploratory prototyper; he is bad at finishing/committing to dates. Treating his work as plannable creates dependencies that bite Wolford (Cedric reassignment) and creates pressure on Greg that he wont meet anyway. Firewalling preserves the option value of Gregs experiments without poisoning forecast hygiene. Aligns directly with Peters confidence-as-contract doctrine — items must have planned dates and confidence; Gregs work does not, therefore it cannot be on the plan.

Additional Context

Directly tied to the Cedric write-off decision (he is functionally Gregs); the Kyle conditional move (Greg-pull is real and structured); and the QBR Intelligence Hub feedback (Product team needs a real voice-of-customer mechanism, not Greg-driven instincts).

Observed Evidence

Direct quote above; reinforced by Cedric-write-off context in same call.

Matching Patterns

Confidence Breakdown

Reasoning Depth Analysis