How to Read a 10-K with AI, Section by Section

To read a 10-K with AI, work through it section by section rather than asking for one summary: start with the business description, then the risk factors, MD&A, the financial statements and their notes, asking the AI focused questions about each and requiring a quote and section reference for every answer. That way you get the speed of AI without losing track of what the filing actually says.

A one-line request to "summarise this 10-K" tends to produce a pleasant overview of the company that tells you little you could not find on its website. The value in a 10-K is in specifics, and specifics need targeted questions.

What is a 10-K?

A 10-K is the annual report that US public companies file with the SEC. It includes audited financial statements and a standard set of sections, called Items, covering the business, risks, management's discussion and other disclosures. The SEC's own guide, How to Read a 10-K, explains the structure for investors. You can find any company's 10-K on SEC EDGAR.

Which sections of a 10-K matter most?

Part III items, such as executive pay, are often incorporated by reference from the proxy statement.

How do you read Item 1, the business section, with AI?

Ask the AI to explain the business model in specific terms:

Using only Item 1 of the attached 10-K, describe how the company makes money: its segments, main products, customer types and any disclosed customer concentration. Quote the relevant passages.

Check for customers that account for a large share of revenue; companies often disclose customers above 10% of revenue. That information feeds directly into understanding dependencies, which we cover in stock supply-chain dependency mapping.

How do you read the risk factors with AI?

Risk factors are long, and many are generic. The useful questions are about specificity and change:

List the risk factors in Item 1A that name a specific customer, supplier, product, regulator or geography. Then compare with last year's 10-K and list any risk factor that is new, removed or materially reworded.

The comparison is the important part. A new risk factor naming a single supplier tells you much more than the twentieth paragraph about general economic conditions. See filing risk-factor change analysis and how to track changes in SEC filings.

How do you read MD&A with AI?

MD&A is where management explains the numbers. Ask about drivers, not just figures:

From Item 7, what reasons does management give for the change in revenue and operating margin? Quote each reason. What does it say about liquidity, capital spending and the coming year? Note any hedging language.

Look for explanations that are vague ("various factors"), new emphasis and topics that receive less attention than last year.

How do you read the financial statements with AI?

AI can extract and compare figures quickly, but this is where errors creep in. Ask precisely:

From the consolidated statements in Item 8, give revenue, gross profit, operating income, net income, operating cash flow and capital expenditure for the last two fiscal years. State units and quote the line items exactly as labelled.

Then check the figures against the statements yourself. Watch for:

• Units. Thousands versus millions.
• Period. Fiscal years that do not match calendar years.
• Basis. The 10-K uses GAAP; adjusted figures usually appear in earnings releases instead.
• Restatements. Prior-year figures may have been revised.

The notes deserve their own questions: revenue recognition, segment information, debt maturities, leases, stock-based compensation and commitments and contingencies.

What should you check in controls and the auditor's report?

Ask whether management or the auditor reports a material weakness in internal control over financial reporting, and whether the audit opinion includes any going-concern paragraph. These are rare but important. A quick question covers it:

Does the 10-K disclose any material weakness in internal control, any going-concern doubt or any change of auditor? Quote the passage or state that none is disclosed.

How do you keep AI answers tied to the source?

Four habits:

1. Ask for quotes and Item numbers with every answer.
2. Tell the AI to say "not found" rather than guess.
3. Spot-check figures that affect your decision.
4. Keep the comparison filing to hand. Change matters more than level.

Tools differ in how well they cite. Uploading the filing to a general assistant works; filings connectors can retrieve and cite automatically. We compare the options in SEC filings in ChatGPT and Claude. Our earlier, shorter piece on analysing a 10-K with AI covers the principles.

How should you read a 10-K for a company you don't own yet?

When you are researching a new idea rather than checking a holding, change the order. Start with Item 1 to make sure you understand the business, then go straight to Item 1A and the notes on revenue and segments before reading management's account in MD&A. Reading the risks before the narrative helps you notice what MD&A plays down.

Useful questions at this stage:

• What would have to go wrong for this business to lose money?
• Which disclosed risks are specific to this company, rather than to every company?
• How much of revenue depends on one customer, product or region?
• Does operating cash flow broadly track reported profit over several years?
• Is anything unusual in the auditor's report or the controls section?

If you still like the idea after that, read last year's 10-K quickly to see what changed, and the most recent 10-Q for anything new since the annual report.

What should you do with what you find?

Turn your reading into a few checkable claims about the company, and write them down. That gives you a thesis you can monitor when the next 10-Q arrives. Our guide to verifying a stock thesis with AI shows how.

Where does Ecomerate fit?

Ecomerate is an autonomous AI research and portfolio-monitoring agent for DIY investors, free to start. Its answers are evidence-backed with sources. Once you own a stock, it monitors your thesis against new SEC filings and earnings, detects changes such as reworded risk factors, maps supplier and customer dependencies through Chains, and sends alerts by iMessage with the passage attached.

This article is for general information and research education only. It is not investment, tax or legal advice, and it is not a recommendation to buy or sell any security. Check material facts against current primary sources before making a decision.