Key Takeaways
- Supply the ratios yourself; compare with 3 competitors
- Ask for bull, base and bear cases, not a price target
- A high P/E means high expectations already priced in
A Great Company Can Still Be a Bad Buy — If the Price Is Wrong
Every investor learns two separate skills. The first is judging a business. The second — the one that quietly decides most returns — is judging the price. A wonderful company bought at a silly price can go nowhere for years, while a merely decent company bought cheaply can do very well.
This guide is the valuation step of our AI-prompts investing cluster. It builds on the earnings step of the cluster and on the general method in How to Use ChatGPT to Analyze Stocks Before Buying: 6 Powerful AI Prompts, and it works best with the figures from those steps in hand. Same ground rule as always: paste in the numbers yourself — the current share price, earnings per share, and any ratios your broker shows you. Never ask an AI to quote you a live price from memory.
Prompt 1: Explain the Ratios in Plain English
"This stock trades at a price-to-earnings ratio of [x], a forward P/E of [x], and a price-to-sales ratio of [x]. Explain what each ratio means in plain English, and what investors paying these multiples are probably expecting from this company."
The P/E ratio is simply the price divided by one year of earnings per share — at a P/E of 20, you are paying $20 for every $1 of annual profit. That framing turns an abstract multiple into a sentence anyone can judge.
Prompt 2: Compare It With Its Real Competitors
"Here are the P/E and price-to-sales ratios for [company] and three competitors: [paste all four]. Compare them. Which is cheapest relative to its earnings, which is most expensive, and what might explain the differences? Use only the figures I gave you."
A P/E of 28 means nothing alone. Next to competitors at 18, 20 and 22, it means the market expects this company to outgrow the pack — which may be true, but now you know exactly what has to go right.
A worked example (hypothetical numbers, real method). Say your stock trades at 28 times earnings while its three closest rivals sit at 18, 20 and 22. The prompt's comparison makes the gap concrete: you are paying roughly 40% more per dollar of profit than the peer average. That doesn't make the stock a sell — fast growers often deserve a premium — but it tells you the burden of proof the company has to meet.
Prompt 3: Build a Valuation Range From Your Own Assumptions
"If this company earns $[x] per share next year, what would the share price be at P/E ratios of 15, 20 and 25? Show the math for each. Then do the same if earnings come in 10% lower than expected."
This is the most useful prompt in the set, because it replaces a vibe with arithmetic. Three multiplications give you a range — and ranges are honest in a way single price targets never are.
Prompt 4: Run the Bull, Base and Bear Cases
"Using the figures I provided, lay out three scenarios for this stock over the next year: a bull case, a base case and a bear case. For each, state the earnings assumption, the P/E you are applying, the resulting price, and the one thing that would have to happen for that scenario to play out."
Notice what this prompt refuses to do: it never asks "will this stock go up?" It asks what has to be true for each outcome — a question an AI can actually help you think through.
Prompt 5: Ask What Is Already Priced In
"Based on this stock's valuation compared with its competitors, describe the level of growth and success that seems to be already reflected in the current price. Then list the disappointments that would hurt this stock most, given those expectations."
Expensive stocks rarely fall because a company does badly. They fall because a company does well, but not well enough for the price that was paid. This prompt names that risk before you're holding it.
What This Process Can't Do
It cannot tell you the stock's true value, because no such single number exists — value depends on future earnings nobody can know today. It also cannot use live data you didn't give it, and it will happily build precise-looking scenarios on top of any wrong figure you paste in. The defence is the habit from the first guide in this cluster, 7 ChatGPT Prompts to Research Stocks Like an Analyst: verify the inputs against the company's actual filings before you trust the outputs.
Series note: Previous: read stock earnings reports with ChatGPT. Next: find and evaluate Canadian dividend stocks. Related: why cheap stocks become value traps.
The Twikup Take
Valuation is where investing stops being storytelling and becomes arithmetic — and arithmetic is exactly where AI help shines, provided you supply the numbers and demand scenarios instead of predictions. Pay for the business you're actually getting, not the one in the headline. Next in the cluster: turning the same lens on dividend stocks, where the trap isn't the price — it's the yield.
Image: Illustrative image.