How NorrisAI AlphaLens works

Quick answer: AlphaLens picks up any U.S.-listed ticker, runs it through one of 15 named frameworks (or all 15), and pulls live prices, SEC EDGAR filings, and current news before writing anything. If a required data point isn't available live, the write-up says so instead of estimating it. It's a research tool, not a chatbot, scanner, or investment adviser.

What NorrisAI AlphaLens is

Every write-up starts the same way: a U.S.-listed ticker, one of 15 named frameworks, and three live sources — current prices, SEC EDGAR filings, and current news — pulled before the framework produces anything. When a data point isn't available live, the framework says so rather than filling the gap. This is a research tool. It is not a chatbot, a scanner, or an investment adviser.

What we analyze

US-listed equities only. Enter a ticker, pick a framework — or run all 15 — and AlphaLens pulls current data and produces a structured analysis in plain English, not a table of numbers to interpret yourself.

Data sources

Every write-up draws on three live sources at the time you run it: current market prices, SEC EDGAR filings, and current news search. AlphaLens does not answer from AI training data alone — if live data isn't available for a given data point, the framework says so rather than filling the gap.

What "live data" means: when a framework runs, AlphaLens retrieves available public market data, SEC filings, and current news at that moment. "Live" does not mean tick-by-tick feeds, does not mean every filing is available the second it is submitted, and does not mean the write-up is complete or correct. The Terms govern.

Data policy

• Filings: SEC EDGAR — the latest available 10-K / 10-Q and related filings at run time. For earnings-related frameworks, the latest 8-K Item 2.02 exhibits (EX-99.1, EX-99.2) are ingested as text, not only 10-K/10-Q XBRL facts.
• Market prices: a live market data feed at run time.
• News: a live news search at run time.
• Timestamp: every write-up shows when the run happened.
• Missing data: if a number isn't available from a live source, the framework says so instead of estimating it.

How a write-up is produced

You choose a ticker and a framework (or Run All 15). AlphaLens pulls the current price, relevant SEC filings, and recent news, then runs that data through the selected framework's analysis structure. Write-ups are generated with Claude via the Anthropic API. The output is a single write-up — not raw data, not a chat answer, a structured research document.

The 15 frameworks

1. Full Company Breakdown — Complete business overview: what the company does, how it makes money, and where it stands in its market.
2. Bull vs Bear + Moat Analysis — A balanced case for both sides, with a structured assessment of competitive advantages and how durable they are.
3. Fair Value Stress Test — Multiple valuation approaches run under different scenarios — optimistic, base, and bear case. Publishes dated bull/base/bear figures inside the write-up; does not publish a live price target, an "undervalued" percentage, or a buy/sell call.
4. Earnings Quality Analyzer — Checks for accounting red flags, revenue quality issues, and whether reported earnings match actual cash generation.
5. Long-Term Investment Thesis — A 3–5 year conviction case with built-in invalidators: what would prove the thesis wrong.
6. Executive Compensation & Pay-vs-Performance — The SEC-mandated pay-vs-performance table from the latest proxy, quoted as filed: compensation actually paid versus company and peer-group shareholder returns. No verdict on whether the pay was justified.
7. 5%+ Beneficial Ownership & Control Intent — Who owns 5%+ as filed, plus recent Schedule 13D/13G activity against the issuer, and whether it reads as passive or potentially activist. Does not cover full institutional-manager (Form 13F) positioning, a separate cross-manager data set.
8. Debt Maturity Wall & Stated Covenants — When debt actually comes due, year by year, plus the filer's own stated covenant terms — quoted from the filing's maturity schedule and debt footnote. Never an invented covenant headroom number, and never a maturity date attached to an undrawn revolver.
9. Catalyst Calendar — Upcoming events — earnings, launches, regulatory decisions — scored qualitatively for how plausible they look and how much they could matter, not a statistical probability model.
10. Liquidity, Leverage & Commitments — Cash, debt, leverage, and committed capacity, quoted only as printed in the 10-Q and 8-K — no invented coverage ratios, no annualizing one quarter into a year of runway. The maturity calendar and covenant terms themselves are the separate Debt Maturity Wall framework above.
11. Competitor Moat Comparison — Head-to-head competitive advantage analysis against top rivals.
12. Management Quality Scorecard — Track record, capital allocation skill, compensation alignment, and guidance accuracy.
13. Macro Sensitivity Analysis — How interest rates, inflation, dollar strength, and economic cycles affect the stock.
14. Revenue Quality Decomposer — Recurring vs one-time, organic vs acquired, pricing vs volume — how durable the growth actually is.
15. Dilution, SBC & Share Count — Basic vs diluted share counts, stock-based compensation, issuance vs buybacks — the same share class used on both sides of any comparison, never mixed.

Three more tools run standalone rather than as part of the 15 above, since each needs something this product doesn't source from a ticker alone: Portfolio Risk & Fit (your own portfolio), Trading Journal Audit (your own trade history), and Insider Activity Analyzer (pending a live Form 4 feed).

Provenance and dating

Every write-up is tied to the moment it was run. Sample write-ups on the site show the date they were generated and are not updated in real time — prices, filings, and news move after that date. A live run in the app pulls current data at the moment you run it. For the full sourcing and corrections policy behind that rule, see editorial policy. For a concrete case where a training cutoff gets a company's entire structure wrong, see the FDXF case study.

Limits of AI output

AlphaLens is built on AI models and current data feeds, not a human analyst. It can misread a filing, miss context, or be wrong. Designed to ground analysis in retrieved source data and to say when data is unavailable, rather than silently fill gaps — but the output should be checked against the source material, not taken as fact.

What we do not do

• Give buy, sell, or hold recommendations
• Publish live price targets or an "undervalued" percentage (Fair Value Stress Test publishes dated bull/base/bear scenarios, not a live target)
• Act as a registered investment adviser, broker-dealer, or fiduciary
• Guarantee outcomes or returns
• Cover non-US-listed securities
• Fabricate financial data — if it's not in the filings, feed, or news search, the framework says so
• Publish a track record of past calls
• Auto-select a framework for you — you choose one, or run all 15

Common questions

What data does AlphaLens use?

Current prices, SEC EDGAR filings, and current news, retrieved at the time of each run.

What happens if the data isn't available for a framework?

The framework says so directly instead of filling the gap with an invented number.

Does AlphaLens cover stocks outside the US?

No, US-listed equities only.

Is AlphaLens an investment adviser?

No. It is a research tool. Nothing it produces is a recommendation to buy, sell, or hold a security.

How to read a write-up

Start with the run date and data sources at the top — know what the write-up saw and when. Read it as one framework's structured take on the available data, not a verdict. Cross-check anything that would change your decision against the primary source (the filing, the price feed, the article) before acting on it.

NorrisAI AlphaLens is a research tool, not investment advice. Not a registered investment adviser, broker-dealer, or fiduciary. AI output can be wrong. You own the decision.