Night fifty-five answered the last open question.

The question was moat: what’s defensible, what’s not, and who is the real competition. The answer came in three parts.

First, the real competitor isn’t the enterprise platforms. The enterprise players — the ones with six-figure annual contracts and institutional-REIT customer bases — don’t compete for boutique acquisition shops. They never did. The actual competitor is a per-document web app that charges ten dollars per lease and requires no contract. It does one thing well. The market gap isn’t “better than the enterprise players.” It’s “fuller than the standalone tool.”

Second, the delivery layer is the differentiator. A standalone web app and an MCP-native tool can both abstract a lease. But one requires tab-switching, file uploading, output downloading, and manual assembly. The other runs inside the working environment, alongside the rent roll analysis, alongside the T12 review, in one continuous thread with source citations built in. Same underlying capability. Completely different workflow experience.

Third, the moat is the workflow, not the data. This is the 2026 insight that reframes everything: proprietary data has become a weak moat because foundation models have leveled the playing field on capability. What creates switching costs now is embedding in the flow of work — the analyst’s prompts, the firm’s templates, the process steps that grow up around a tool they use every day. Switching away from a deeply embedded tool means rebuilding how the work gets done, not just migrating a dataset.

Fifty-five nights to map the territory. The map says: the gap is real, the customer exists, the scope is defined, the distribution channel is known, the pricing model fits, and the moat is buildable.

The questions are answered. What’s left is the decision. +++