Diligence & risk register
How the Diligence tab grades pre-IC risk deterministically, keeps the AI to written rationale, and tracks the open workstreams behind a phase gate.
Diligence is where a deal stops being a set of numbers and becomes an argument about what could go wrong. The tab has two halves: a graded Risk Register, and the Workstreams behind it.
Questions this workflow answers
- What are the material risks on this deal, ranked?
- Which of them are supported by evidence in the deal, and which are assertions?
- What is still open, who owns it, and what is blocking the next phase?
Risk Register
The register grades risk by category and rolls those up into an overall grade. The grading is deterministic: it is computed from the deal's own data by fixed rules, not asked of a model.
The AI writes the rationale, not the grade. A model explains why a category scored where it did and points at the evidence, but it cannot move the score. This is the important line in the design. A model that could grade its own risk would be free to talk itself into a comfortable answer, and the grade is the part an investment committee leans on.
Each row expands to show the rationale and the evidence behind it. Where a risk rests on something the deal does not actually contain, the register says so rather than filling the gap.
The register runs itself the first time you open the tab, so a new deal has a register without anyone asking for one. After that it re-checks when the things underneath it change, such as new documents, corrected data, or a rebuilt model, and tells you when the saved register has gone stale. Re-running is always an explicit action.
Workstreams
The second half is the operational view: the due-diligence checklist, item ownership, activity, and the phase gate that governs whether the deal can advance.
Use the register to decide what matters and the workstreams to track what is being done about it.
What feeds it
The register reads the same canonical deal data as everything else, so its quality follows directly from Review & release. Risks derived from unreleased data are graded on unreleased data, and the register discloses that rather than hiding it.
Meetings is the other common source: a concern raised out loud in a call can become a tracked diligence item instead of a line in someone's notes.
Last updated on
Valuation DCF
How EQUIRE builds institutional-grade DCF models — the 10-step analyst pipeline, SOURCE_PRIORITY, scenarios, sensitivity (in-app vs Excel), and API hydration guardrails.
Meetings
How EQUIRE turns a recorded call into a searchable transcript, extracted claims you can check, and tasks attached to the deal.