How to Read an Earnings Report With AI
Last updated June 2026
Short answer
To read an earnings report with AI, start with the parts that move a stock: the headline numbers (revenue and EPS), the guidance, the segment detail, the management commentary and tone, and how the stock actually reacted. Ask a general assistant like ChatGPT or Perplexity to summarize the release and explain each part in plain language and against expectations, then verify the specific figures against the company’s own filing, because AI can state numbers wrong. A connected assistant like Walnut can relate the print to a holding you actually own, framed against the S&P 500. Walnut is not an investment adviser.
An earnings report can feel like a wall of numbers, and the numbers that get the headline are often not the ones that move the stock. AI is genuinely good at cutting through it: it can summarize a release in seconds, define the jargon, and point you at what changed. What it cannot do is be trusted blindly on the exact figures, so the workflow that matters is summarize fast, then verify. This guide walks through the parts of an earnings report in the order they matter, shows how to use AI on each one, and is honest about where you still have to check the source yourself.
What an earnings report is
An earnings report is the quarterly (or annual) update a public company gives on how the business did. It comes in two pieces: a press release with the headline numbers and management commentary, and a formal SEC filing (the 10-Q each quarter, the 10-K for the full year) with the detail. Most companies also hold an earnings call where executives take analyst questions. The report answers one question in a lot of detail: did the business grow, make money, and where is it headed?
The trap for a new reader is treating the numbers in isolation. Markets price in a forecast before the release, so what matters is the result against expectations and against the company’s own prior outlook. That framing is exactly the kind of context AI is good at supplying quickly, as long as you verify the figures it leans on.
Read the parts in the order that matters
The fastest way to read a report is to go part by part, and to ask AI a specific question at each step rather than “is this good.” Here is each part, what it is, and what to look for.
Headline numbers (revenue and EPS)
The top-line figures in the press release: revenue (total sales) and earnings per share (profit divided by shares). These are what the headlines and the stock react to first.
- What to look for: Whether each figure came in above, at, or below what analysts expected, and how it compares to the same quarter a year ago.
Guidance (the outlook)
Management’s forecast for the next quarter or full year, usually a range for revenue and earnings. It is the company’s own view of what comes next, not a past result.
- What to look for: Whether the outlook is raised, held, or cut versus prior guidance, since a strong quarter with weak guidance often moves the stock down.
Segment and detail lines
The breakdown beneath the headline: revenue by business unit or geography, margins, and any key operating metrics the company reports (users, units, subscribers).
- What to look for: Which segment drove or dragged the result, whether margins expanded or compressed, and whether growth is broad or concentrated in one line.
Management commentary and tone
The CEO and CFO remarks in the release and on the earnings call: what they credit, what they flag as a risk, and how confident they sound about demand.
- What to look for: The language around demand, pricing, and costs, and whether the tone matches the numbers or quietly hedges them.
The stock reaction
How the share price moves after the print, in the after-hours session and the next day. It reflects the result against expectations, not the result on its own.
- What to look for: Whether the move fits the numbers, since a beat can still fall on soft guidance and a miss can rise on a better outlook or a one-time charge.
Notice that guidance and the stock reaction sit at the end but often explain the most. A beat can fall on soft guidance; a miss can rise on a better outlook. If the price move does not match the headline numbers, the answer is almost always in the outlook or in one segment, and that is the thing to ask AI to dig into.
How to summarize a release with AI, fast
General assistants are the right first tool here. They turn a dense release into something you can scan in a minute, and they explain the terms as they go.
- ChatGPT or Claude. Paste the press release text (or link it in a browsing mode) and ask for a summary organized by revenue and EPS versus expectations, guidance, top segment drivers, and management tone. Claude tends to hold the thread across a long release; ChatGPT is a fast, flexible all-rounder. Neither sees live prices, so ask it to explain the numbers, not to quote them.
- Perplexity Finance. Good for a cited quick answer: ask what a company reported and whether it beat or missed, and it links the sources and shows quotes and recent news inline. That makes it easy to trace a figure back, which is exactly what you want before acting.
- Ask expectation-framed questions. “Did revenue beat consensus?” “Was guidance raised or cut?” “What drove the segment result?” get you further than a vague “summarize this,” because they force the model to compare against a benchmark.
For the wider toolkit, see the best AI investing tools roundup, and for a deeper method, how to analyze earnings with AI.
Then verify, because AI gets numbers wrong
This is the step people skip, and it is the one that matters. AI is reliable for structure and explanation and unreliable for exact figures: it can transpose a number, attribute growth to the wrong segment, or use a stale consensus estimate. None of that shows up as an error; it reads as confident prose.
- Trace the headline figures to the filing. Confirm revenue, EPS, and the guidance range against the company’s own press release or SEC filing, not the summary.
- Check the estimate it compared to. A “beat” is only meaningful against the right consensus number; ask which estimate the AI used and sanity-check it.
- Read the actual guidance language. Outlook is where the stock reaction usually hides, and it is easy for a summary to soften or drop a hedge.
Summarize fast, verify slow. The AI saves you the reading time so you can spend it judging the outlook and checking the numbers that would actually change your view.
At a glance
| Part of the report | What to look for |
|---|---|
| Headline numbers (revenue and EPS) | Whether each figure came in above, at, or below what analysts expected, and how it compares to the same quarter a year ago |
| Guidance (the outlook) | Whether the outlook is raised, held, or cut versus prior guidance, since a strong quarter with weak guidance often moves the stock down |
| Segment and detail lines | Which segment drove or dragged the result, whether margins expanded or compressed, and whether growth is broad or concentrated in one line |
| Management commentary and tone | The language around demand, pricing, and costs, and whether the tone matches the numbers or quietly hedges them |
| The stock reaction | Whether the move fits the numbers, since a beat can still fall on soft guidance and a miss can rise on a better outlook or a one-time charge |
Relate the print to what you actually own
A general assistant explains the earnings report in the abstract. It does not know whether you hold the stock, how much, or how it sits against the rest of your portfolio, unless you tell it. That last mile is where a connected assistant helps.
To be upfront, since this is our site: Walnut is an AI financial assistant whose chatbot is grounded in your real holdings. It connects your existing brokerage through SnapTrade (read-only by default) and lets you ask how a fresh earnings print relates to a position you actually own, by talking through Claude, ChatGPT, or a built-in assistant, with each holding framed against the S&P 500. So after you have read and verified the release, you can ask a question like “how does this quarter change the picture for a name I hold” against your real position rather than a hypothetical.
Walnut is not a data terminal and does not replace reading the filing: it sits on top of your broker, leans on web and price data, and frames returns as window returns because broker feeds rarely pass cost basis. It is read-only by default, every trade needs your approval, and Walnut is not an investment adviser. For the research workflow around a single name, see how to research a stock with AI.
The bottom line
Reading an earnings report with AI is a two-step habit: summarize fast, then verify. Ask a general assistant like ChatGPT, Claude, or Perplexity Finance to break the release into the headline numbers, the guidance, the segment detail, the management tone, and the stock reaction, and to explain each against expectations. Then trace the specific figures back to the company’s own filing, because AI states numbers with confidence it does not always earn. A connected assistant like Walnut can relate the print to a holding you actually own, framed against the S&P 500. Read fast, verify carefully, and keep the decision yours. Walnut is not an investment adviser.
Try Walnut on top of your broker
Walnut connects any major US broker in a few clicks, then lets you ask how an earnings print relates to what you hold through Claude, ChatGPT, or its built-in AI, with each position framed against the S&P 500. Read-only by default; you approve every trade.
FAQ
How do I read an earnings report with AI?
Paste or link the earnings release into an assistant like ChatGPT, Claude, or Perplexity and ask it to summarize the headline numbers, the guidance, the segment detail, and the management tone. Then ask what came in above or below expectations and what moved the stock. Verify the specific figures against the actual filing, because AI can state numbers wrong. Walnut is not an investment adviser.
What are the most important numbers in an earnings report?
Revenue, earnings per share, and guidance. Revenue is total sales, EPS is profit per share, and guidance is the company’s own forecast for next quarter or the year. The result matters less on its own than against what analysts expected. A strong quarter with weak guidance can still send the stock down, which is why the outlook often matters most.
What is a beat versus a miss?
A beat is when a reported figure comes in above the analyst consensus estimate; a miss is when it comes in below. Companies are measured against expectations, not just against last year, so a growing company can still miss if it grew less than forecast. AI can quickly tell you whether revenue and EPS beat or missed, but confirm the estimate it used against a reliable source.
Why did the stock drop after a good earnings report?
Usually because the result was already expected, or because the guidance disappointed. Markets price in a forecast before the print, so a beat that merely meets high expectations can fall, and soft guidance can outweigh a strong quarter. Ask an AI assistant to compare the numbers to consensus and to summarize the outlook, then look at what the guidance said versus the prior quarter.
Can ChatGPT explain an earnings report?
Yes. ChatGPT is strong at summarizing a release in plain language, defining terms, and walking through what a segment result or a guidance change means. On its own it cannot see live prices or your brokerage, and it can state figures incorrectly, so treat it as a fast explainer and verify any specific number against the actual report before acting on it.
How do I use AI to relate earnings to a stock I own?
General assistants do not know what you hold unless you tell them. A connected assistant like Walnut links your brokerage through SnapTrade (read-only by default) so you can ask how a fresh earnings print relates to a position you actually own, with the holding framed against the S&P 500. You still do your own verification, and any trade is yours to approve.
What should I read first in an earnings report?
Start with the press release headline: revenue and EPS versus expectations, then the guidance. Those three usually explain most of the stock reaction. From there, read the segment detail to see what drove the result and the management commentary for tone. AI can surface all four in seconds, which lets you spend your time judging the outlook rather than hunting for the numbers.
Is AI accurate when summarizing earnings?
AI is reliable for structure and explanation but not always for exact figures. It can transpose a number, mix up a segment, or use a stale consensus estimate. Use it to understand what a release says and why it matters, then check the specific revenue, EPS, and guidance numbers against the company’s own filing or press release before you rely on them.
What is guidance and why does it matter so much?
Guidance is management’s forecast for upcoming quarters, usually a revenue and earnings range. It matters because a stock prices the future, not the past, so a raised outlook can lift shares even after a soft quarter, and a cut outlook can sink them after a strong one. When reading with AI, always ask whether guidance was raised, held, or lowered versus before.
Where do I find a company’s earnings report?
The primary sources are the company’s investor relations page and its filings with the SEC (the 10-Q or 10-K and the 8-K press release). Financial data sites and Perplexity Finance also surface the numbers with citations. Whatever tool summarizes it for you, trace the key figures back to the original filing so you are reading the company’s own words, not a paraphrase.
Can AI tell me whether to buy or sell after earnings?
It can explain the result and frame the trade-offs, but a trustworthy assistant will not tell you to buy or sell. Walnut is informational and is not an investment adviser: it helps you read the print, relate it to what you own, and frame holdings against the S&P 500, but the decision and any trade are yours. Be wary of any tool promising a sure call from a single quarter.
Walnut is informational and is not an investment adviser. App features, pricing, and availability change; verify current details on each provider's site before deciding. Nothing on this page is a recommendation to buy, sell, or hold any security or to use any particular product.