A diagnostic and prescriptive layer that reads your ledger via OAuth and never writes back. Connect QuickBooks, Intacct, NetSuite, or Xero in an afternoon — then let the AI tell you what went wrong and what to do about it.
We don't replace your accounting software. We sit on top, read-only by default, and add the diagnostic and prescriptive layer that finance leadership has been doing in Sheets for the last decade.
Close-automation tools answer the first question, faster. We answer two more.
The four modes are not four products. They are four ways of asking the same engine the same question. Pick the mode that matches the question; the math is the same.
What makes the suite prescriptive rather than a glorified spreadsheet is the AI Planning Layer — seven distinct AI surfaces, all calibrated by your operational driver framework. The model can see what your business has actually done before, and it's bounded to propose only what's plausible inside that envelope.
A 5% price increase your team has held three times before scores very differently from a 25% increase you've never attempted. The calibration is the moat — the more the platform sees, the more credible its proposed paths become.
"Why did the AI propose this scenario?" is a question we expected, so we made it answerable. Every Prism call lands in ai_planning_calls: the surface, the model, the prompt payload, the response payload, the triggering user, the timestamp. Operators can see exactly what the AI was told and why it returned what it returned.
The "AI-in" parts — generation, narration, NL parsing — are AI-driven. The "AI-out" parts — the feasibility scoring, the envelope clamping, the simulator itself — are deterministic. The boundary is explicit. The score is reproducible even when the model is not.
Connect your books in an afternoon. See the diagnostics and the first scenarios on your real data, the same week. No implementation fees, no six-month rollout, no SOW.