The 10-K is the single most important document in fundamental investing. It's the company's annual report to the SEC — audited financials, risk disclosures, management's narrative, and the fine print that often moves stocks 5% overnight.
Yet most investors never read it. They rely on summaries, broker reports, or simply the stock price to make decisions. That's a gap this guide aims to close.
Whether you're a retail investor building your own research process or a professional looking to optimize your workflow, here's how to read a 10-K filing effectively — and where artificial intelligence can help.
What's Actually in a 10-K Filing
The SEC requires public companies to file a comprehensive annual report. Here's the section-by-section breakdown:
| Section | What It Contains | Why It Matters |
|---|---|---|
| Item 1: Business | Company overview, products, services, markets | Understand what the company actually does |
| Item 1A: Risk Factors | Material risks — regulatory, competitive, operational | Identify what could go wrong |
| Item 2: Properties | Physical assets, facilities, leased spaces | Assess operational scale |
| Item 3: Legal Proceedings | Pending lawsuits, regulatory matters | Flag potential liabilities |
| Item 5: Market for Common Stock | Share price history, dividends, holders | Understand shareholder returns |
| Item 6: Selected Financial Data | 5-year summary of key metrics | Spot trends over time |
| Item 7: MD&A | Management's Discussion & Analysis | The narrative behind the numbers |
| Item 8: Financial Statements | Balance sheet, income statement, cash flow | The hard data |
| Item 9A: Controls & Procedures | Internal control effectiveness | Assess reporting reliability |
| Item 10-14 | Corporate governance, exhibits, filings | Governance structure |
What Analysts Actually Care About
Professional investors don't read every section with equal intensity. Here's where they focus:
1. Risk Factors (Item 1A)
This is where you'll find what keeps management up at night. The SEC requires companies to disclose "material risks" that could adversely affect the business.
What to look for:
- New risk factors added since last year (often signal emerging threats)
- Risk factors that have been removed (signifies resolved issues)
- Specificity of language — vague risks are less actionable than specific ones
2. Management Discussion & Analysis (Item 7)
The MD&A is where management explains the numbers in their own words. It's not audited, which means it's their spin on the business.
What to look for:
- Changes in tone compared to prior quarters (confidence vs. caution)
- Forward-looking statements about revenue, margins, and capital allocation
- Explanations for significant variances (why did revenue drop? why did margins expand?)
- Segment-level detail — how are different business units performing?
3. Revenue and Segment Breakdown
The financial statements show what happened. The segment breakdowns in the notes show where it happened.
What to look for:
- Revenue by segment (not just total revenue)
- Gross margin by segment (Some segments may be more profitable than others)
- Geographic breakdown (How much revenue comes from China? From Europe?)
- Concentration risk (Who are the top customers?)
4. Notes to Financial Statements
These are the footnotes — and they're anything but trivial. This is where you'll find:
- Revenue recognition policies (how they count sales)
- Litigation contingencies (potential liabilities)
- Stock-based compensation details (dilution)
- Leasing commitments (future obligations)
How AI Handles Each Section
This is where the game changes. AI doesn't replace judgment — but it can do the first pass on every page.
Speed and Comprehension
An AI agent can process a full 10-K in 14-30 seconds, compared to 4-6 hours for a human analyst. That means you can:
- Cover 10x more names in the same time
- Get alerts on new filings within minutes of publication
- Maintain consistent coverage across your entire watchlist
Structural Extraction
AI excels at extracting structured data from unstructured text. It can:
- Pull all risk factors into a searchable database
- Extract revenue by segment across multiple years for comparison
- Identify changes in management tone by comparing word choice
- Flag forward-looking statements for human review
What AI Gets Right
1. Data extraction: Numeric data, segment breakdowns, and explicit statements are extracted with high accuracy.
2. Pattern recognition: AI identifies patterns across multiple filings — a company adding "China export controls" as a risk factor in 2023, 2024, and 2025 shows consistent exposure.
3. Comparison: AI can instantly compare a company's 10-K against its 10-Ks from prior years, flagging what's changed.
4. Consistency: Every company gets the same analytical framework, making peer comparisons more meaningful.
What AI Still Gets Wrong
Full transparency matters. Here's where human judgment remains essential:
1. Tone and Subtext
AI can extract words, but it struggles with what's implied. When a CFO says "we remain confident in our long-term strategy" during a period of declining revenue, that's not a positive signal — it's a deflection. Human readers understand this. AI often doesn't.
2. Industry Context
A 15% revenue decline at a growth-stage SaaS company means something different than at a mature retailer. AI can describe the numbers but lacks the industry knowledge to contextualize them.
3. Management Credibility
Reading a 10-K from a company with a history of accounting restatements requires extra skepticism. AI doesn't have a "trust meter" that adjusts based on past behavior.
4. Uncertainty Calibration
Companies use specific language to signal confidence or caution. "We believe" vs. "We expect" vs. "We project" carries different weight. AI often treats all forward-looking statements as equivalent.
5. The "So What"
AI can tell you that revenue was $8.1 billion, up 12% YoY. It struggles to explain why that matters compared to consensus, what it implies for the valuation, or how it fits the broader investment thesis.
The Practical Workflow in 2026
The most effective approach combines AI speed with human judgment:
1. AI First Pass — Let AI extract the data, flag key sections, and identify changes from prior periods. This replaces the 4-6 hour manual reading with a 2-minute review of AI-generated highlights.
2. Human Deep Dive — Focus your time on the sections that require qualitative judgment: MD&A tone, risk factor changes, and industry context.
3. Synthesis — Build your investment thesis from the AI-extracted data and your human analysis of what's missing from the numbers.
Where SignalPress Fits
SignalPress is designed for investors who want the AI first pass without building their own pipeline.
The system pulls directly from SEC EDGAR — no paid data subscriptions required — and generates a structured brief with:
- A narrative thesis (not just bullet points)
- Key metrics with YoY and sequential context
- Risk factors ranked by materiality
- A directional conviction signal with reasoning
Start with 3 free briefs on any ticker you're researching. The AI reads the latest 10-K, extracts the signals, and generates the brief in 14 seconds. You'll see exactly what it gets right — and where your judgment still matters.
Want to see how AI handles the full filing analysis? Read our deep dive: [How AI Reads SEC EDGAR Filings in 14 Seconds](/blog/how-ai-reads-sec-filings)
Comparing research tools? Here's our analysis: [AI Investment Research Tools: What Actually Works for Fund Managers](/blog/ai-investment-research-tools)