01
UNDERSTAND
How does this business actually work?
Map the customer, product, revenue model, costs and dependencies. Identify what has to go right for an investor to earn a return.
Investment diligence skill · design partner preview
The report is the output.
The investigation is the product.
A rigorous AI investigation skill for venture capital, private equity, and private credit teams. It tests management's claims against the underlying records, challenges competing explanations, and traces the consequences to potential investor returns.
WATCH THE INVESTIGATION 03:12 · On-screen text (no audio)
Built for investment teams./Find the few questions that matter./Test them against the record./Make the downside explicit.
HOW THE INVESTIGATION WORKS
From a data room to an investment conclusion you can defend.
The skill does not start by writing a report. It first works out what would make or break the investment, then investigates the questions that could change the decision. Each conclusion is tested against the records, not simply repeated from a pitch deck.
START HEREYour existing AI workspace + the company's data room
02
PRIORITIZE
Which unanswered question could change the deal?
Choose the most consequential uncertainty, such as whether new sales are repeatable or retention is sustainable, instead of checking everything with equal weight.
03
INVESTIGATE + CHALLENGE
What do the original records actually show?
Rebuild the relevant figures from source files, test credible rival explanations, and separate what is proven from what still needs evidence.
04
FOLLOW THROUGH
What changes for the investor?
Trace the result into growth, margins, cash needs, future dilution and potential returns. Investigate the next material question, or state what cannot yet be resolved.
Repeat for the next decision-changing question. No unsupported certainty.
The output: a source-linked investment investigation
See the sample PDF
A clear thesis, decisive evidence, alternative explanations, scenarios, risks, open questions and conditions that would change the conclusion.
EXAMPLE / COLDTHREAD
One deal. Several numbers worth challenging.
The synthetic ColdThread data room presents a convincing growth story. The investigation goes deeper than the headline metrics: it recalculates customer retention, compares operating cost classifications, and tests whether the sales plan can support the proposed valuation.
ColdThread and its operating records are fictional. Figures illustrate the research method and are not real company performance.
95% → 89.1%Claimed vs. full-cohort gross retentionThe reported figure omits smaller customers and a major exit.
66% → 43%Reported vs. delivery-payroll-adjusted gross marginIncluding support and hardware operations staff changes the economic picture.
$1.75M vs. $0.78M2027 new-customer ARR plan vs. latest actualThe investment hinges on whether selling productivity can accelerate.
EXPLORE THE SAMPLE
See the analysis behind the conclusion.
Everything below is available without sharing a real deal. Take a look at the report, then tell us what you would want tested in your own underwriting workflow.
01 / READ · PDF
ColdThread investment report
A 13-page synthetic investment committee example with operating analysis, scenario charts, risks, and decision conditions.
Open the PDF ↗01Management: retention looks healthy
02Rebuild customer-by-customer cohorts
03Check alternative churn explanations
04Translate to financing and returns
02 / WATCH · VIDEO, 03:12
ColdThread investigation explainer
See how the skill chooses consequential questions, verifies management metrics and traces different company outcomes to investor returns.
Watch the explainer ↑DESIGN PARTNER TRIAL
Try it on one of your deals.
Interested in using the skill inside your team's existing AI environment?
Let's walk through a sample and hear what your process needs.