Step 1: Ask for the business first
Before any number, get the plain picture. What does the company sell, to whom, and how does it earn money? If the answer is vague, the rest won't help.
A good prompt is short: "In five sentences, how does this company make money, and which segment is biggest?" Then check the answer against the company's own annual report or investor page.
Step 2: Pull the core numbers
Next come revenue, margins, free cash flow and debt. Ask for the trend over several years, not one quarter. A single good year says little. Five years say more.
A report such as the AI stock analysis in Peakwise puts these figures next to a short comment on each. The comment saves time. The figures are still yours to verify.
- Revenue growth over several years
- Operating margin and net margin
- Free cash flow, the cash left after investment
- Debt compared with equity or cash flow
Step 3: Look at value as a range
A price only means something against what the business might be worth. Ask the AI for a fair value as a range, with the method named. Low, mid and high. A single target price hides how uncertain the estimate is.
Our guide to fair value explains why a range is more honest than a point.
Step 4: Ask for the risks and the bear case
AI tends to agree with the tone of your question. Ask only about the upside and you get upside. So ask the opposite on purpose: "What are the three strongest arguments against owning this stock?"
Good risks are specific. "Competition" is filler. "One customer makes up 30% of revenue" is a risk you can check.
Step 5: Check the chart and the news
The chart shows how the market has treated the stock. The news shows what changed. Read both before you form a view. For charts, start with how to read a stock chart.
What AI is good at, and where it slips
AI is good at summarizing long documents, turning tables into plain language, and laying out arguments on both sides. It is quick, and it doesn't get bored reading a 200 page filing.
It slips on exact numbers, on recent events a model hasn't seen, and on anything that needs a source it can't name. It can also invent a figure that looks right. That is the main reason to check.
- Good: summaries, explanations, comparing arguments, spotting what is missing
- Risky: exact figures, dates, ratios from memory
- Risky: anything after the model's training data ends
- Risky: predictions. Nobody, human or AI, knows next year's price
How to check an AI answer
Pick the two or three numbers the conclusion depends on. Find them in a primary source: the annual report, the quarterly filing or the company's investor page. In the US, filings are public on SEC EDGAR.
Tools that search the web for each answer and link the sources make this easier. The investment chatbot in Peakwise does that, so you can open a source and compare. If a tool gives no sources at all, treat its numbers as leads, not facts.
A quick check (illustrative)
Say an AI answer states that a company's free cash flow was 4 billion dollars last year. Open the cash flow statement. Operating cash flow minus capital spending should land near 4 billion. If it doesn't, ask the AI where its number came from. Two minutes, one number.
What the result is, and what it is not
An AI analysis is a starting point for your own research. It is not advice, and it can't know your goals or how much risk you can stand. Write your own conclusion in two sentences. If you can't, you haven't finished.
Frequently asked questions
Can AI analyze stocks?
Yes, as a research helper. It can summarize financials, news and charts and lay out arguments for and against. It can't predict prices, and it can state wrong figures, so check the numbers that matter.
How do I analyze a stock with AI?
Ask for the business model, pull the core numbers over several years, look at value as a range, ask for the bear case, then read the chart and news. Verify the key figures in a primary source.
Is AI stock analysis reliable?
It is reliable for structure and explanation, less so for exact figures. Treat numbers as leads until you have opened the source.
What should I ask an AI about a stock?
Start with how the company makes money, then ask for revenue, margins and debt trends, a fair value range with its method, the strongest arguments against the stock, and what could change in the next year.
Does Peakwise tell me what to buy?
No. Peakwise is a research tool. Its analysis shows data, a fair value range and risks. The decision stays with you.