An analyst opens a new deal folder and loses the first two hours rebuilding a checklist she's used on the last three deals. Follow these 6 steps to build a reusable framework, brief your materials in one pass, and turn verified findings into a first draft of your memo.
An analyst opens a new deal folder on a Tuesday afternoon: a data room link, a management deck, and a deadline for the investment committee memo. Before she can ask a single real question about the business, she spends the first two hours rebuilding the same diligence checklist she used on the last three deals — pulling it from a shared drive, adjusting it for this sector, re-reading her own notes to remember which red flags mattered last time. The work that actually matters, judging whether the numbers hold up and the story is credible, hasn't started yet.
This is the quiet cost of due diligence at most funds and research desks: not the analysis itself, but the setup that has to happen before the analysis can begin. Every deal restarts from a blank framework, even when the questions being asked are 80% the same as the last one.
In this guide, we'll walk through 6 steps to run AI-assisted due diligence: building a framework you can reuse across deals, briefing it on a target's materials, surfacing gaps and unverified claims, and turning verified findings into a first draft of your memo.
What You'll Need
- A due diligence checklist or framework you already use, even an informal one (industry-specific or not)
- The target company's available materials — financials, a data room, management decks, or public filings
- A place to keep notes and open questions across the life of the deal
How to Use AI for Due Diligence: 6 Steps
Step 1: Build a Reusable Framework Instead of Starting From Scratch Each Time
Before touching a new deal, take the checklist you already use — even if it currently lives in your head or a scattered set of old memos — and turn it into a structured list of the questions you always ask: revenue quality, customer concentration, management background, competitive position, whatever applies to your sector. The goal isn't a new framework; it's making your existing one reusable so you're not reconstructing it from memory on deal five.
Try this with Noumi: "Here's the rough checklist I use for SaaS diligence: revenue quality, churn drivers, customer concentration, management track record, competitive moat. Turn this into a structured framework with sub-questions under each category that I can reuse on future deals in this sector."
Tip: Save this framework at the project level rather than rewriting it inside each deal's notes — that way it carries forward automatically the next time you open a similar deal.
Step 2: Brief Your Workspace on the Deal Before You Ask Any Questions
Once the framework exists, hand over what you have on the specific target: the data room contents, the management deck, prior call notes, whatever materials are available at this stage. Do this in one pass rather than answering questions about the company piecemeal as they come up — the context needs to be in place before the framework can be run against it.
Try this with Noumi: "I've uploaded the data room financials, the management deck, and my notes from last week's call with the CFO for [Target Co]. Read through these and confirm what's covered before I run my diligence framework against them."
Step 3: Run Your Framework's Questions Against the Materials
With the framework and the deal materials both in place, work through the checklist category by category and ask what the materials actually support versus what they're silent on. This step is where most of the time savings happen — not because the judgment call gets automated, but because pulling the relevant figure or clause out of a 200-page data room stops being the bottleneck to asking the next question.
Try this with Noumi: "Using the SaaS diligence framework, go through the customer concentration section: what percentage of revenue comes from the top 10 accounts, and does the data room disclose any contract terms tied to renewal risk?"
Example output:
- Top 10 accounts: 34% of ARR (disclosed in data room, tab 4)
- Renewal terms: not disclosed for the top 3 accounts — flagged as open question
- Churn commentary in the deck references "low churn" without a specific number — flagged as unsupported claim
Step 4: Flag Claims That Aren't Backed by the Materials You Have
Management decks are written to tell a good story, and not every claim in one comes with a number attached. As you work through the framework, keep a running list of statements that sound confident but aren't backed by anything verifiable in the materials — "strong retention," "best-in-class NPS," "no material customer risk" — so they become explicit questions for management rather than assumptions that slip into your memo unchallenged.
Tip: Treat this list as a required deliverable of the diligence process, not a side note. A short list of "claims we couldn't verify" is often more useful to an investment committee than a longer list of claims that were.
Step 5: Keep a Running List of Open Questions and Follow-Ups Across the Deal
Diligence rarely finishes in one sitting — new documents arrive, management calls answer some questions and raise others, and a deal can sit for weeks between active phases. Keep the open-question list attached to the deal itself rather than scattered across email threads and old chat messages, so picking the work back up in three weeks doesn't mean re-reading everything to remember where you left off. Choosing tools built for this kind of ongoing research work matters here specifically because the context needs to persist across sessions, not just within a single afternoon of work.
Try this with Noumi: "Add these three items to the [Target Co] open-questions list: renewal terms for the top 3 accounts, the specific NPS number behind the 'best-in-class' claim, and the management team's plan for the VP of Sales vacancy. Flag any of these that are still unanswered when I check back next week."
Step 6: Turn Verified Findings Into a First Draft of Your Memo
Once the framework is worked through and the open-question list is current, the memo itself should mostly assemble from work you've already done rather than requiring a separate writing pass from a blank page. Ask for a draft organized around your standard memo structure, built strictly from what the diligence actually found — verified figures, flagged claims, and outstanding questions — then review it against the source materials before it goes to the committee.
Try this with Noumi: "Draft the diligence section of the investment memo for [Target Co] using the findings from the customer concentration, churn, and management sections we've worked through. Keep the flagged and unverified items in a separate 'open questions' section rather than folding them into the main narrative."
Pro Tips for Faster, More Consistent Diligence
- Sector-specific frameworks beat one generic checklist. A SaaS framework and a manufacturing framework should ask different questions; maintaining separate versions pays off after the second or third deal in each category.
- Don't skip the framework step to save time on a "simple" deal. The deals that feel simple at the start are often the ones where a skipped category turns into the surprise later.
- Review the open-questions list before every management call. Walking into a call with last week's unanswered items already listed turns the call into a follow-up session instead of a repeat of the first meeting.
- Keep the source citation with every figure. A number without a page or tab reference in the data room is much harder to defend three weeks later when someone on the committee asks where it came from.
Frequently Asked Questions
Getting Started
Due diligence doesn't get faster by working longer hours on the same rebuild-from-scratch process — it gets faster when the framework and the context from each deal carry forward instead of resetting every time. Start with one sector-specific checklist this week, and build the habit of briefing your materials in one pass before you start asking questions against them.
The open-questions list is usually the single highest-leverage habit here: it turns diligence from something you finish once into something you can pick back up accurately after any gap, without re-reading everything to remember where you left off.
If your deal flow means picking up the same coverage context weeks or months after the last active phase, Noumi is built to hold that context between sessions so your framework, your open questions, and your prior findings are all still there when you need them. Try Noumi →
