The record is real.
An independent reviewer rebuilt the hash chain from the spec — not trusting our code — and verified every record across all eleven workspaces. Verdict: no fabrication found.
A firm that sells proof should hold itself to the same bar. So we built our own companies to test it — ten fictional businesses across five industries, plus one deliberately oversized stress case — and ran each end to end through the live, real-time system. None of these are customers. They are test simulations on real software, with every step recorded the way a real client’s would be.
Each industry got two test companies: a mature one with a full, well-kept policy set, and a scrappy one with thin, messy documents — because real governance has to survive both. Every one was run through the same pipeline: documents in, rules cleaned and mapped, controls signed off by three separate roles, simulations against new laws and partners, and a full hash-chained audit trail.
Meridian Holdings is a fictional conglomerate we made deliberately monstrous: 1,408 pages of policy across four volumes. We pushed it through to see where the system bends under real scale — and it mapped 11,125 live controls while keeping its audit trail intact end to end.
Scale wasn’t the only test. Across the fleet we ran 94 simulations — new laws, new territories, partner companies connected with evidence-only handoffs — and built 50 reports, all with their hashes embedded so a tampered one is detectable.
We didn’t grade our own homework. Three separate independent audits re-checked the work from different angles — rebuilding the hash chain from scratch, proving the system can’t bypass its own audit gate, and recomputing every published statistic against the raw records.
An independent reviewer rebuilt the hash chain from the spec — not trusting our code — and verified every record across all eleven workspaces. Verdict: no fabrication found.
A second reviewer proved the system physically cannot skip its audit gate, then recomputed every verdict against the sealed engine — zero mismatches.
A third reviewer recomputed every statistic on this page from the raw snapshots, and corrected anything that read rosier than the records. What you see here is what actually ran.
Each test company was run end to end through the live system. We are publishing the full account of each — what it faced, what the engine caught, and what the record proved. New studies land here as they are written.