根拠となるエビデンスを自ら保持するデューデリジェンス

Definition and scope

Continuous diligence is a structured part of an investment process, with an identifiable question, an evidence base and an accountable handoff. Its purpose is not to produce more paperwork. It is to make the reasoning behind a recommendation understandable to someone who did not conduct the original research. The scope should reflect the investment mandate, the asset or entity under review and the decision being requested. A useful record separates facts, assertions, assumptions and unresolved questions. It also states the limits of the work, including areas that require specialist review. This guide describes a practical process rather than a promise about investment outcomes.

Reuben AI is decision infrastructure for private capital. It brings diligence, research and the investment committee into one connected record, so every decision is consistent, evidence-based and traceable to its source. Judgement stays human. The process becomes disciplined.

How it works, step by step

1. Set the scope and responsibility

   Identify whether the review concerns a company or asset, vehicle, fund or manager. Record the jurisdiction and investment structure because they affect the questions to ask. Assign workstream owners and identify where legal, tax, technical or other specialist advice is needed.

2. Register evidence and its limits

   Connect each material finding to the document or research that supports it. Record whether a statement is self-reported, checked against another source or supported by independent evidence. Keep conflicting material visible rather than choosing the document that best supports the preferred conclusion.

3. Resolve gaps through accountable questions

   Turn each material uncertainty into a question with an owner, a requested source and a decision consequence. A question can remain open if the committee knowingly accepts the uncertainty. It should not disappear merely because a report is ready for circulation.

4. Continue review after the decision

   Carry conditions and thesis assumptions into ownership. When new information changes a material finding, preserve the earlier reasoning and document the revised interpretation. Continuous diligence means an accountable review process, not a promise that every external event is detected automatically.

Review checklist

• The scope and requested decision are explicit, and the relevant mandate is attached.
• Material claims have supporting sources, with limitations and conflicting evidence visible.
• Facts, management assertions and analytical assumptions are distinguishable to the reader.
• Open questions have owners and consequences for the recommendation, rather than unexplained status labels.
• The receiving reviewer can find the rationale, accepted risks and any required specialist advice.
• Conditions and follow-up responsibilities carry into the next stage without losing their original context.

Common mistakes

• Confusing document collection with completed diligence
• Closing unresolved questions without recording the reason
• Treating the approval memo as the end of the evidence record

These mistakes have a shared cause: the summary becomes detached from the work behind it. Before a handoff, ask a colleague to trace a material conclusion back to its source and explain what would change the recommendation. If they cannot do that, clarify the reasoning before polishing the presentation. A concise qualification is more useful than an unsupported conclusion, and an explicit open question is more useful than an empty field that looks complete.

Worked example: illustrative

Illustrative example, not a client case study: consider a proposed investment in a manager-led fund. The team begins by writing the decision question and linking the initial thesis to the material available. A management presentation supports the proposed opportunity, but a supporting document leaves an important assumption uncertain. The analyst records the disagreement rather than smoothing it into a positive summary. The workstream owner requests the underlying evidence and explains how the answer could change the recommendation. When the response arrives, the reviewer checks whether it addresses the actual question, records its limitations and updates the finding. The committee sees both the revised view and what remains open. If it proceeds subject to a condition, that condition moves into ownership with a named owner. A later review compares new information against the same assumption. The example illustrates the chain from evidence to judgement without assuming that the investment succeeds or inventing commercial results.

Evidence and interpretation

A reference guide is useful only if a reader can distinguish what is known from what is inferred. Give each important claim a source, explain the source’s relevance and keep any qualification beside the conclusion it affects. A source can be genuine and still be insufficient for the decision. For example, a statement about a broad industry may not apply to a particular customer segment, contractual arrangement or jurisdiction. The reviewer should be able to follow the reasoning without guessing which passage mattered. Where sources conflict, explain the conflict and the next question rather than silently selecting a preferred version.

Ownership and review discipline

Name the person responsible for the next action and explain what would count as a satisfactory response. A request to obtain more information is not complete until someone has considered what that information means. The review should distinguish a closed question from an accepted risk and from a matter outside scope. Preserve the reason for each status so another colleague can continue the work. If the scope changes, record the change explicitly. This makes the process usable when responsibilities move between research, diligence, committee preparation and portfolio oversight, without suggesting that a standard template replaces professional judgement.

Boundaries and handling sensitive material

Use only information the team is authorised to access and share. Confirm the appropriate audience before circulating a source or a summary derived from it, particularly when external advisers or platform partners are involved. Keep the decision record separate from any suggestion that software provides legal, tax or investment advice. Teams remain responsible for their governance arrangements and the decisions they make. Deployment and integration requirements should be scoped with Reuben AI rather than assumed from a workflow description. A useful evaluation starts with the team’s evidence and review requirements, then checks how the proposed setup supports them.

Related questions

How does this work connect to the next decision? Use the other topic guides to follow the evidence through research, diligence, committee preparation and ownership review.

• Research and screening
• One record for the investment lifecycle
• Investment committee
• Investment memo

Questions in this topic

• Best AI due diligence software
• Embedding diligence in an investor platform
• Automated due diligence research pipeline
• Diligence for platform clients
• Diligence at every level
• Diligence by asset class and jurisdiction
• How AI diligence avoids hallucinated findings
• How long AI-assisted diligence takes
• Acquisition due diligence
• Best due diligence software for corporate development
• Running several acquisition diligence processes
• Search fund acquisition diligence checklist
• How family offices run direct deal diligence
• Family office co-investment diligence
• Corporate venture capital due diligence
• What a diligence vendor security review covers
• What an auditable diligence trail contains
• Operational due diligence review
• Manager level due diligence
• How to read a private company data room

Frequently asked questions

What is the purpose of continuous diligence?

Continuous diligence is the repeated testing of an investment thesis as evidence changes, rather than a report that ends at approval. Reuben AI keeps fully sourced findings, unresolved questions and committee reasoning connected through ownership.

Does Reuben AI make the investment decision?

No. People assess the evidence, challenge the recommendation and make the decision. Reuben AI connects the material used in that process; it does not give investment advice.

How should unresolved questions be handled?

Keep them visible beside the recommendation, with an owner, the evidence requested and the consequence for the decision. Distinguish a question that has been answered from a risk the decision makers knowingly accept.

What makes a source useful for review?

The source must support the particular claim being made, within its actual scope. Record the relevant document context, distinguish assertions from checked findings and explain any contradictory evidence.

When should the record be reviewed again?

Review it when new evidence could change a material assumption, when a decision condition requires follow-up or when the organisation’s review process calls for it. Preserve the earlier rationale when recording a changed view.

Where is data stored?

Data residency is available in APAC, EMEA or the Americas, with the region agreed during onboarding.

What is an evidence gap in investment diligence?

An evidence gap is a missing, outdated or insufficiently supported fact needed to evaluate a material claim. It is not proof that the claim is false.

Who should own unresolved diligence questions?

Assign a named workstream owner to obtain and assess the evidence, with an accountable deal lead responsible for escalation to the committee.

When should missing evidence delay an IC decision?

Delay when the missing evidence could change eligibility or a material conclusion and the committee cannot responsibly bound the uncertainty. Record the reason and the evidence needed to resume.

How should low-priority gaps be treated?

Keep them visible with a rationale for deferral and a review date. Low priority should mean limited decision relevance, not silent removal.

Should a company finding automatically become a fund finding?

No. Explain the exposure, ownership relationship and relevance to the fund before drawing a fund-level conclusion.

How should duplicate findings be handled?

Reference the same underlying finding where appropriate and identify overlapping exposure, rather than counting copied findings as independent evidence.

What belongs beside a rolled-up conclusion?

Include the entity, source reference, document date or version, material qualifications and the reviewer responsible for the interpretation.

How should missing underlying data appear?

State the coverage boundary and the unavailable evidence explicitly. Do not present a partial roll-up as a complete portfolio view.

How do we distinguish an investment red flag from an unresolved evidence gap?

A red flag is an adverse finding or credible warning that needs investigation; an evidence gap is insufficient support for a conclusion. A gap may reveal a red flag, but absence alone is not proof of wrongdoing.

How should a resolved red flag stay visible to the committee?

Keep the original concern, investigation evidence, resolution rationale, reviewer and date in the decision record. Show any residual risk and required follow-up alongside the resolution.

Who authorises restarting a paused review?

The authority specified by the investment team’s governance process should confirm the required evidence has been reviewed and explain the restart decision.

Should every concern stop the process?

No. Assess materiality, source reliability and mandate implications. Escalate unresolved material concerns rather than treating every minor discrepancy as disqualifying.