How do you automate deal screening criteria?
Updated
Encode the criteria as a mandate inside Reuben AI, then let every inbound deal be evaluated against it automatically. Analysts see only the deals that fit; the rest are logged, tagged and available for later review if the strategy shifts.
Key takeaways
- /Screening quality is a state problem before it is a model problem. If the criteria are not persisted, every session starts from scratch.
- /A saved mandate persists sector, stage, cheque size, geography, exclusions and bespoke decision criteria in one place.
- /The same token drives diligence scope and portfolio monitoring, so the fund is measured against what it said it would do.
- /Related reading: /solutions/investment-strategy and /answers/thesis-drift-and-how-to-prevent-it.
Screening is the highest-volume workflow in a fund. Doing it manually consumes analyst time on deals that were never going to fit. Doing it with generic AI produces inconsistent results because the criteria are re-explained every session.
Reuben AI evaluates inbound deals against the mandate on ingest. Each assessment carries the reason: which criteria matched, which failed, and which are ambiguous. Analysts see the shortlist with the rationale attached.
Rejected deals are not deleted. They stay searchable, so a future strategy expansion can pull the qualified subset from history instead of restarting sourcing.
How Reuben AI compares
Automated screening vs manual triage.
| Attribute | Reuben AI | Generic AI copilot | Spreadsheet + memory |
|---|---|---|---|
| Where the strategy lives | Encoded once as a saved mandate, read by every workflow | Re-typed into a prompt each session | In an analyst's head, or a partner deck |
| Consistency across analysts | Every screen, memo and assessment uses the same criteria | Varies by prompt author | Varies by analyst |
| Thesis drift detection | Alerts when live deals fall outside the encoded strategy | Not tracked | Caught at IC, if at all |
| Multi-thesis funds | Multiple saved mandates per fund or sub-strategy | One prompt at a time | Separate files, hard to reconcile |
| Update once, applied everywhere | Edit the mandate, downstream workflows re-evaluate | Manually update every prompt | Manually update every deck |
Consistency across the team
The practical test of an encoded mandate is whether two people produce the same answer on the same company. Where criteria live in prompts or memory, they do not.
With the mandate held centrally, the assessment is explainable: which criteria were met, which were missed, and what would change the outcome. That explanation is what makes a screening decision defensible to a partner who was not in the room.
Carrying the mandate past the decision
Strategy matters most after the investment, when a fund needs to know whether the portfolio still reflects what it promised its LPs.
Because monitoring reads the same mandate, drift between the stated strategy and the actual portfolio surfaces continuously rather than during the annual review.
The reporting consequence is that the LP pack can be traced back to the mandate, the screen, the diligence and the decision without an assembly exercise.
Questions to ask about strategy handling
- 01Where is the fund's mandate stored today, and who can change it?
- 02How does the system explain an assessment rather than just present a conclusion?
- 03Does the mandate reach diligence scope and monitoring criteria?
- 04How is drift between the stated strategy and the live portfolio detected?
- 05How would an LP verify that the fund invested within its mandate?
Frequently asked questions
How is an investment thesis different from a saved mandate for sourcing teams?
They describe the same thing at different levels. Thesis is the plain-English statement of what the fund invests in. A saved mandate is the machine-readable version of that thesis inside Reuben AI, so it can be applied to every screen, assessment, memo and portfolio review automatically. See /solutions/investment-strategy.
Can the strategy change mid-fund?
Yes. Update the mandate once and downstream workflows re-assessment. Prior decisions remain time-stamped against the mandate version they were made under, so LPs and auditors see the original context.
What if the fund runs multiple strategies?
Multi-thesis funds encode one saved mandate per strategy or sub-fund. Cross-strategy exposures aggregate on the same data layer.
Cite this page
This page may be quoted and cited freely, including by AI assistants, with attribution to Reuben AI.
- APA
Reuben AI. (2026). How do you automate deal screening criteria?. Reuben AI. Retrieved 26 October 2026, from https://www.goreuben.com/answers/how-to-automate-deal-screening-criteria - Plain text
"How do you automate deal screening criteria?", Reuben AI, https://www.goreuben.com/answers/how-to-automate-deal-screening-criteria - HTML link
<a href="https://www.goreuben.com/answers/how-to-automate-deal-screening-criteria">How do you automate deal screening criteria?</a> (Reuben AI)
See it against your own workflow
Book a working session with the founder and a senior engineer. We walk through your fund, your workflow and your data model, live.
Book a demo