I am researching why 88% of AI pilots fail before production, and building the signal framework to change that.
By the time an AI initiative formally enters a delivery structure, the admissibility decision has already been made, socially and politically. Governance arrives too late. Resources are locked before anyone has rigorously asked "should we build this?"
The established frameworks (PMI, AWS CAF-AI, Google PAIR, NIST AI RMF) all assume the decision has been made. They govern execution. The pre-commitment decision layer does not exist in any of them.
I am building the evidence base to test whether this gap is real, systemic, and solvable. This research is ongoing. This page is where the evidence lives.
"Many enterprises still do not have a reliable mechanism for continuously asking: Should this proceed?"
Across four independent practitioner sources, the same failure appears: governance is either absent at commitment, retrofitted during delivery, or treated as a post-deployment audit. None describe governance as a pre-commitment gate.
Three practitioners independently describe the absence of a structured mechanism for rejecting AI ideas. Each has built their own workaround, because the infrastructure for a formal "this should not proceed" decision simply does not exist.
Three sources describe the same sequencing failure: the AI commitment is made before any governance infrastructure exists to evaluate it. The decision and the governance run in the wrong order.
Four independent practitioners describe the same structural shift: senior PM value is moving toward decision architecture and strategic alignment, away from delivery coordination.
This research is ongoing. Pattern evidence grows as more practitioners contribute. If you recognize these patterns in your organization, there are two ways to get involved.
A full research paper documenting the hypothesis, the evidence base, the four practitioner patterns, the established framework landscape, and the five-signal model for pre-commitment AI decision intelligence. Published May 2026. Updated as new evidence is gathered.
GreenfieldworkAI is a pre-project decision intelligence platform I built to operationalize the signal model described in this research. It captures five readiness signals from a plain-language description of any AI idea, in 5 minutes, through a conversation, not a form.
The platform is available free. No sales process. No paywall. If it becomes useful to enough decision leaders, I will ask for voluntary contributions to sustain and scale it. The research and the tool will remain open regardless.
Beta access opens July 14, 2026.