How to research a stock using AI

Researching one stock well is a sequence of questions answered in the right order. AI can answer most of them faster than you could alone — but only if you keep the sequence and verify the output. To research a stock with AI, move from business → quality → value → price, using AI to gather evidence at each step and keeping the final judgment for yourself. Here's the walkthrough.

1. Understand how the company makes money

Start with the simplest, most important question: how does this business earn a dollar? Ask AI to read the latest annual report and explain the revenue model, the main customers, and the cost structure in plain English. If the explanation is convoluted, that's information — some businesses are genuinely too hard to value, and recognizing that early saves you from false precision later.

2. Pull the long-term numbers

Have AI gather 5–10 years of the figures that reveal business quality:

• Revenue growth and its consistency
• Gross, operating, and free-cash-flow margins
• Return on invested capital versus the cost of capital
• Debt levels and how they've trended

A decade of data tells you whether you're looking at a durable compounder or a single good year dressed up as a trend.

3. Find the moat — or confirm there isn't one

Ask AI to identify the source of durable advantage: switching costs, network effects, cost advantages, or intangible assets. Then make it defend the answer with evidence from margins and market share. If neither you nor the AI can name which moat the company has, assume it has none and price accordingly.

4. Value it conservatively

Now the math. Ask AI for two things:

• A discounted cash flow on conservative assumptions — modest growth, normalized margins, an honest discount rate — producing a value range, not a point estimate.
• A reverse-DCF: what growth does today's price already bake in? If the market is assuming the business doubles its earnings forever, you don't need a model to know the price is demanding.

5. Check the margin of safety

Compare the current price to the conservative end of your value range. The margin of safety is the discount between them — and the discount you require should be wider for harder-to- forecast businesses. This is the step that turns research into a decision: does the price clear the bar this specific business demands?

6. Stress-test before you act

Before any conclusion, ask AI the question that matters most: "What would make this a bad investment?" Get the bear case, the fragile assumptions, the risks hidden in the footnotes. A thesis that survives a genuine attack is one you can hold when the price falls.

7. Verify, then decide

Trace the key numbers — the margin, the growth rate, the debt figure — back to the actual filing. AI can misread or fabricate, and a valuation built on a wrong input is worse than no valuation. Once the evidence checks out, the decision is yours to make, in plain words you could defend to a skeptic.

The shortcut

This whole sequence is exactly what Claremont Street automates. Enter a ticker and the 167-point framework runs it end to end — filings read, quality scored, conservative value range built, margin of safety tested, bear case included — so you get a structured, evidence-backed report in place of an afternoon of prompting.

---

This analysis is for informational and educational purposes only and is not investment advice. Claremont Street is not a registered investment advisor. Do your own research.

Related reading:

• How to do investment research using AI
• What is AI investing?
• How we think about moats
• Research any ticker with the 167-point framework →