How to analyze a CIM

A good CIM review is not a line-by-line hunt for errors. It is a structured effort to understand the seller’s argument, identify the assumptions carrying that argument, and assign each important claim to the right form of verification.

Start with the investment question. If you do not know what decision the document must inform, every chart can look equally important. Then move from the broad narrative to claims, evidence, and management questions.

A seven-step buyer-side CIM review

1. Write the deal thesis in one sentence. Note what must be true for the opportunity to work: durable customer demand, credible expansion, defensible product differentiation, or another specific driver.
2. Map the seller’s narrative. Record how the company describes its history, market, customer, product, performance, and growth. Treat this as an argument to understand before evaluating it.
3. Build a claim ledger. Capture the exact statement, page number, category, time period, and the evidence needed. Preserve the original wording; paraphrase only in a separate field.
4. Sort by verification path. Financial claims belong in the model and quality-of-earnings work. Contract claims require agreements. Public-history claims can be compared with dated web records, filings, releases, and other contemporaneous sources.
5. Prioritize material claims. Focus on statements that affect the thesis, valuation, risk allocation, or management credibility. A harmless copy change should not crowd out an uncertain customer-concentration assumption.
6. Form neutral questions. Ask for definitions, dates, scope, and primary support. A question that permits an ordinary explanation will usually produce better diligence than an accusation.
7. Record the conclusion and its limits. Separate supported evidence, missing evidence, inconsistencies, and items still awaiting management response.

What to extract from the document

The most useful ledger entries are specific enough to investigate. “Strong product” is an opinion; “the company has served enterprise infrastructure customers since 2017” contains a customer segment and date. Useful categories include:

• Founding, launch, expansion, and operating-history dates
• Customer count, segment, tenure, concentration, and logo references
• Product launches, capabilities, integrations, and market positioning
• Locations, employees, service areas, and geographic reach
• Certifications, partnerships, awards, and performance representations
• Pricing, packaging, contract structure, and go-to-market changes

Do not force every item through the same evidence source. Website archives can illuminate public positioning and timing; they cannot establish audited revenue, private contract status, or the intent behind a change.

How to use AI without outsourcing judgment

Model assistance can accelerate extraction and organize candidate questions, but it should not be the evidence. Require every extracted claim to point back to the exact document passage. Require every historical observation to point to an original dated source. Let an analyst approve the claims researched and the findings exported.

A reliable workflow also preserves negative states: unsupported within scope, inconclusive, out of scope, and awaiting primary evidence. Those labels prevent a weak archive result from being mistaken for proof that something never happened.

Limitations to state in the memo

A CIM reflects the seller’s selected presentation, not the entire business record. Public archives are incomplete. A capture date shows when an archive stored a page, not necessarily when the page was first published. Pages may be missing, blocked, or rendered incorrectly. Historical copy can change for ordinary strategic or editorial reasons.