A junior analyst can use AI to turn financial statements, management presentations and call notes into a first draft of an investment memo. By the time the analyst opens the spreadsheet, there may already be an explanation for why revenue grew and a sensible-looking list of risks.
I can see the appeal. Pulling information out of dozens of documents takes time. It also gives an analyst repeated exposure to the discrepancies that can change a decision. A number in the presentation covers a different period from the accounts. A customer who sounds committed turns out to have made one purchase. You develop a feel for what matters by working through enough cases and hearing how an experienced colleague reads the same facts.
Give analysts a question to own
Suppose a software company says most of its new contracts come from referrals. An AI summary might list that as evidence that customers recommend the product and that sales can grow without a big rise in marketing costs. The analyst needs to find out what the company counts as a referral.
Start with the underlying sales records. Did those introductions come from a broad base of customers or from one reseller? Speak with a customer and ask how they found the product and why they bought it. A call with the sales team might reveal that some of the apparent referrals came from a promotion that has since ended. Each answer changes what the analyst should investigate next.
Ask the analyst to explain how the result affects the investment case, which evidence would change their view, and what they would check before committing money. That discussion builds skills the analyst can carry into the next deal.
Make time for practice
AI can assemble documents, flag mismatches and suggest questions. Managers can use the time it saves to give junior colleagues ownership of a few claims that matter. Let them trace a number to the source, lead part of a customer call and present their recommendation before hearing the senior team’s view.
After the decision, return to the memo. Did the reseller keep sending customers? Did the pipeline turn into orders? An analyst who can compare their forecast with what happened has a concrete basis for improving their judgment. A senior investor can help by explaining where their own reading was off.
If an analyst spends less of the morning gathering numbers, give them part of the afternoon to follow up on the one number they cannot explain. Over several deals, those conversations and checks are how someone learns which question to ask before the next memo is written.
Adapted and expanded for IJK Capital Partners from Is AI bad for humanity? on Kenny Ng’s personal blog.