# Land-and-Expand in Enterprise AI Land-and-expand is the dominant go-to-market for enterprise AI: enter with a cheap pilot, prove value, expand to production pricing. The gap between land and expand is where most companies die. The land is easy. Blue-chip companies run pilots constantly. A $46-60k pilot with a 3-month timeline and an exit clause is a rounding error on an R&D budget. Logos look great on a pitch deck. The expand is existential. Moving from pilot to production means: regulatory validation (6-12 months in pharma), IT security review, procurement cycle, integration with production systems, operator training, and demonstrating ROI with controlled baselines. Most pilots never expand. The logo stays on the deck. The revenue doesn't. The diligence test: break down revenue into new customers vs. expansion of existing. If all revenue is new logos, the expand motion is unproven. If even one customer has increased spend after initial pilot, it's partially validated. These are radically different scenarios for the same revenue number. Pharma is the extreme case. Pre-GxP pilots look like production but aren't. GxP validation adds 6-12 months and $200k+ per deployment. Those Pfizer and GSK logos might be constrained advisory pilots that can never reach production without massive additional investment. The pricing question: if your pilot ACV is $50k and your financial model assumes $1.8M per customer at scale, you need 36x expansion. Without a single data point showing any expansion, that's a hypothesis, not a business model. Related: [[Industrial AI Unit Economics]], [[Incumbent Bundling Risk]], [[Industrial AI MOC]] --- Tags: #deeptech #investing