Key takeaways
- Three readings, kept apart: thesis fit, evidence quality and diligence readiness.
- Every criterion shows the rule, the observed value, the rationale and the source it came from.
- Analysts can override any criterion, and the override needs a written reason.
- A reviewer can reproduce why a deal landed where it did in the queue.
Planned feature: not in Delsivo yet, and details may change before it ships.
What a thesis fit report shows
The report sits on the deal's Brief tab. At the top are three readings. Below them is a row per criterion from your investment thesis rubric.
| Criterion | Rule | Value found | Result | Source |
|---|---|---|---|---|
| Stage | Pre-seed or seed | Seed | Met | Deck v1, slide 14 |
| Geography | HQ in Europe or India | Lisbon, Portugal | Met | Application form |
| Round size | $500k to $3M | Not stated | Unknown | Becomes a question |
| Sector | B2B software | Consumer marketplace | Mismatch | Deck v1, slide 3 |
| Paid usage (preference, weight 25) | Scored 0 to 10 | 6 | Scored | Deck v1, slide 9 and rationale |
Why three readings and not one score
One number hides the reason behind it. A deal can fit your thesis well and still rest on claims nobody has checked. Another can have excellent paperwork and sit outside your strategy. Blending them makes both cases look average.
- Thesis fit: how the deal matches your hard rules and weighted preferences.
- Evidence quality: how much of what matters is backed by a cited source, counting weak extraction and unresolved claims against it.
- Diligence readiness: what is still missing before the deal can move to the next stage.
Deck quality, a separate score of how clearly the deck makes its case, is kept apart from all three so a polished deck doesn't stand in for fit.
Overrides and the reasons behind them
- 1An analyst disagrees with a criterion, for example the sector call on a marketplace that sells to businesses.
- 2They change the result and write the reason.
- 3The report shows the original result, the new one, who changed it and why.
- 4The override is recorded in the deal's audit history and carries into later briefs.
Overrides without a reason aren't accepted. That keeps the report honest for the partner who reads it next week.
Re-running a report
When the founder sends a new deck, answers a request, or you publish a new rubric version, you can re-run screening. Each run is kept with the rubric version, the evidence snapshot, and the model and prompt version used. Comparing two runs shows exactly what changed and why, instead of a score that drifted without explanation.
What the report never does
Those decisions sit with your analysts and partners, as described on thesis screening.
- It never moves a deal to another stage or passes on it.
- It never sends anything to the founder. Questions from Unknown criteria wait for a person to approve them.
- It is never shown to founders. They see a separately configured status, not ratings.