The Deployment Gap

Domain: Deep-Tech (cross-cutting)  ·  Dominant constraint: Institutional commitment

A technology that works in a pilot has cleared the easy test. In large-scale, regulated systems, the hard test is everything that happens afterward: whether the organization will reassign ownership, rewrite contracts, re-train operators, re-architect its data, and make a commitment it cannot easily reverse. The distance between a tool that works and an institution willing to be changed by it is the Deployment Gap.

Demos test tools. Deployments require institutional commitment, contractual redesign, and operational ownership. Because a pilot is deliberately built to sit on the near side of the gap — time-bound, lightly contracted, run by exceptional people, shielded from live operations — its success predicts very little about what happens on the far side. The more carefully a pilot is protected, the less it tells anyone about deployment.

What the gap is made of

Five barriers do the work, and none of them is the technology:

  • Failure intolerance — on a system that cannot go down, a model that is only mostly right is a liability no one will sign for.
  • The data reality — production data is fragmented and messy; the model that was accurate on a clean extract degrades on the real feed.
  • The OT boundary — a sandbox on the IT network is not the control environment; crossing into operational technology is a separate, harder project.
  • The ownership handoff — pilots are owned by innovation teams; platforms must be owned by operations, and the handoff rarely happens cleanly.
  • Contracts and the regulator — scaling means multi-year commitments, explicit liability, and, for a regulated utility, a cost-recovery case; a pilot avoids all three.

Key insight

Pilots test tools. Deployments require institutions to change themselves. Technology that cannot cross that gap does not fail in the lab — it fails at the boundary between innovation and institution.

Read the analysis

Why Utility AI Pilots Die in Production applies the Deployment Gap to AI in the power sector — the pilot graveyard, the five barriers, and what it actually takes to cross.