Condensed co-investment memorandum
Short by design — a check on the lead, not a parallel full IDD.
- Reviewed by
- Partner and IC
- Defensibility
- Every claim cites a source span; length discipline enforced by workflow design.
Co-investment · Workflow
A condensed memorandum on a sponsor-led co-investment — with an explicit section on where you disagree with the lead and a recommended side-letter position if the lead's terms are accepted as written.
Closed-system workflow. Source-backed outputs. Versioned review.
Engagement console
Source intake through review state
Source intake
Workflow stage
disagreement register
Stage running · citations required
Evidence anchors
Open risks
Lead thesis variance
severity · high · cited
Draft artifact
Co-invest memo · v1 draft
schema valid · 47 claims cited
Review state
Partner review
Awaiting challenge before lock
Co-investments arrive with the lead investor's memorandum, GP deck, and sponsor model. Teams often treat the lead's thesis as baseline and add a short note. Disagreements are buried in narrative or left unstated. The instrument is a check on the lead — not a parallel full underwriting — but silence on a point of disagreement is treated as agreement, and that choice belongs to the principal.
Built for senior finance teams who need committee-ready artifacts — not chat transcripts.
Use caseCo-investment alongside a trusted lead
Moment of needAllocation arrives with the lead's memorandum and GP deck
OutputCondensed co-investment memorandum
Use caseIndependent read on sponsor-led co-invest
Moment of needThe question is whether the lead's thesis holds against an independent read
OutputDisagreement register with citations on both sides
Use caseFast co-invest decision
Moment of needCommittee needs a short, cited read — not a full primary IDD
OutputSide-letter recommendation if lead terms are accepted
Use caseCheck on lead underwriting quality
Moment of needCo-invest terms differ from the lead's primary fund terms
OutputExplicit where-we-disagree section
Source materials enter the engagement as a controlled register — scanned, classified, and scoped before the workflow runs.
Source register
Pending admission · scanner gate
| Source | Description | Status |
|---|---|---|
| Lead investor's memorandum. | 01 | |
| GP deck. | 02 | |
| Sponsor model and assumptions sheet. | 03 |
A versioned pipeline — each stage records what happened, what you can review, and what output it produces.
Stage pipeline
Topological DAG · runtime enforced
01
Lead investor memorandum, GP deck, and sponsor model enter the engagement envelope.
02
lead_thesis_review extracts claims, figures, and risks the lead surfaced — each anchored to a source span.
03
disagreement_register enumerates where an independent read of GP deck and model contradicts or qualifies the lead thesis.
04
Recommended side-letter position composed if the lead's terms are accepted as written.
05
synthesise composes the short memorandum, disagreement section, and side-letter recommendation.
06
Signed version locked with audit trail.
Committee-ready artifacts — each with a named reviewer and a defensibility standard.
Short by design — a check on the lead, not a parallel full IDD.
Explicit section with citations on both the lead thesis and the independent read.
Recommended position if the lead's terms are accepted as written.
Co-investment review refuses to re-underwrite from scratch or to assert a disagreement the GP deck does not support. The disagreement register carries dual citations. Partner review is recorded before lock.
Diligence OS supports investment judgment. It does not make investment decisions, provide investment advice, or guarantee risk detection.
Prompt-based tools produce inconsistent outputs with weak provenance. This workflow runs a structured process with an admitted source envelope and a versioned review path.
| Dimension | Generic AI | Diligence OS |
|---|---|---|
| Starting point | Chat-first | Workflow-first |
| Process | User-defined prompts | Versioned process |
| Provenance | Weak provenance | Source-span citations |
| Across documents | One file at a time | The full set reconciled; contradictions surfaced |
| Review | Hard to review | Partner review states |
| Outputs | Inconsistent outputs | Repeatable artifacts |
| Memory | No durable workflow memory | Append-only audit trail |
| Committee use | Difficult to defend | Committee-ready outputs |
| Lead check | Summarize the lead memo | Thesis extract + disagreement register with dual citations |
Starting point
Generic AIChat-first
Diligence OSWorkflow-first
Process
Generic AIUser-defined prompts
Diligence OSVersioned process
Provenance
Generic AIWeak provenance
Diligence OSSource-span citations
Across documents
Generic AIOne file at a time
Diligence OSThe full set reconciled; contradictions surfaced
Review
Generic AIHard to review
Diligence OSPartner review states
Outputs
Generic AIInconsistent outputs
Diligence OSRepeatable artifacts
Memory
Generic AINo durable workflow memory
Diligence OSAppend-only audit trail
Committee use
Generic AIDifficult to defend
Diligence OSCommittee-ready outputs
Lead check
Generic AISummarize the lead memo
Diligence OSThesis extract + disagreement register with dual citations
Institutional outcomes — not productivity slogans.
Closed by construction — each engagement runs inside a dedicated chamber with an append-only audit trail.
Adjacent diligence processes in the same controlled system.
Diligence OS is introduced through a controlled evaluation process. No public signup. No self-serve access.