Skip to content

Tessera

Every number, proven — not promised.

Tessera is attested natural-language analytics for regulated finance. Ask a question in plain English over a governed General-Ledger warehouse and get back not just a number, but the evidence it is right: the exact executed SQL, an independent runtime verdict, and a signed, auditor-verifiable receipt.

The problem this solves

Finance already has chatbots. It doesn't have trust. Every text-to-SQL assistant proves itself offline — "95% accurate on a benchmark." In a regulated report that means 5% is wrong and you don't know which 5%, and the cost of that 5% is a restatement or an audit finding — not a shrug. So finance teams either distrust the tool or pay an analyst to re-check every number, which kills self-service.

Tessera removes the blocker: it proves each answer, at runtime, and hands you a receipt your auditors verify offline — without trusting the system that produced it.

Why it's not just another chatbot

Generic copilot Text-to-SQL assistant Tessera
Runs real SQL on governed data ✗ ✓ ✓
Proves this answer correct, at runtime ✗ ✗ (offline only) ✓ per answer
Catches named finance mistakes ✗ ✗ ✓ 8 failure classes
Auditor-verifiable evidence ✗ ✗ ✓ signed receipt
Fails honest (no number vs. a wrong one) ✗ partial ✓ WARN, never fabricates
Runs fully air-gapped rarely rarely ✓

The differentiator is structural: an independent verifier (re-running the model's own SQL to "check" it is circular and worthless), a per-answer runtime verdict, and a cryptographic receipt.

What it honestly is not

Tessera proves a number reconciles to the certified metric definition — not metaphysical truth. If the semantic layer is defined wrong, agent and verifier agree on a wrong-but-consistent number; governing those definitions is a human responsibility (and building them with the customer is the real consulting work). It is for enumerable, structured analytics (GL / financial metrics), not open-ended document reasoning.

Start here