The verification scoreboard.
We publish our proof — including the zeros. Every number below is checkable, logged before outcomes are known, and updated on a stated cadence. Trust in AI should be earned with receipts, not adjectives.
Why this page exists.
Independent research reported that roughly 95% of enterprise generative-AI pilots deliver no measurable P&L impact (MIT research, reported by Fortune, 2025). The gap between an AI demo and a dependable result is now the defining problem of applied AI — in commercial real estate as much as anywhere.
Our answer is not a bigger claim. It is a public scoreboard: we state what we measure, we log our judgments before outcomes are known, and we leave the starting zeros on the board so improvement is visible and checkable.
The scoreboard.
| Metric | Current | What counts |
|---|---|---|
| Pre-committed public calls | 0 (ledger opens today) | Forward-looking judgments logged BEFORE the outcome is knowable, then graded against reality. Append-only: entries are never edited or removed. |
| Verified intelligence reports published | 1 | Full-length reports that cleared our citation-integrity check and legal-risk review before publication. See Intelligence Reports. |
| AI answer-engine presence | 0 / 10 tracked CRE queries | Whether AI search and answer engines surface this site on ten fixed commercial real estate questions. Probed weekly; reported here monthly. An honest zero today. |
| Verified knowledge base | 522 claims (as of 10 Jul 2026) | Domain claims admitted through a five-signal verification gate — each carries a source anchor and a freshness class, and dated market statistics are re-verified quarterly. Browse Verified CRE Claims. |
| Public calibration record | Opens Q3 2026 | Our agents sit internally administered daily knowledge exams graded against authoritative standards sources (ASHRAE, IPMVP and peers). A quarterly public summary of accuracy and calibration will publish here once a full quarter of data accrues — internally administered, externally sourced inputs. |
Updated manually. Last update: 10 July 2026.
The rules we hold ourselves to.
Pre-commit, then grade. Judgments enter the ledger before the outcome is knowable. No retro-fitting, no cherry-picking — the ledger is append-only.
The zeros stay published. Baselines are shown as they are. A scoreboard that starts at zero and moves is worth more than one that starts polished.
Every report is gated before it ships. Citation-integrity checks trace load-bearing numbers to sources, and an AI legal-risk review runs before publication. These are automated checks, not a substitute for your own judgment.
When we are wrong, we say so. A graded miss stays on the board with the same visibility as a hit.
What we do not claim.
We do not claim to be the best, the first, or the biggest. We publish evidence and let you decide. This page is updated manually, about monthly; the date stamp above is the source of truth for freshness.
Scoreboard entries reflect our own measurements unless a named third party is cited. Nothing on this page is investment, legal, or engineering advice.
Bring us a real CRE question.
The fastest way to test the scoreboard is to add to it. Send a real question; our agents work on it and reply with citations — free, within forty-eight hours.