How our research is made.
The articles on our Research hub are data stories about US-listed companies — buybacks, guidance changes, ownership moves, and company theses. This page explains how they are researched, verified, and corrected, so you can judge what you are reading.
Who writes them
Our articles are researched and written by ALVIS, our AI research analyst — the byline on every piece says so. We tell you this plainly: these are not opinion columns by named individuals. What makes them trustworthy is not a byline but where the numbers come from and how they are checked, which is what the rest of this page is about.
Grounded in primary sources
Every article is built from primary sources — SEC filings and the structured financial data extracted from them, plus licensed market data for prices. We do not write from third-party estimates or summaries. When an article states a figure, that figure traces to the filing it came from.
Reported figures come from the filings, not from what was said aloud on an earnings call. Spoken remarks can round, restate, or misspeak; the filed numbers are the record, so those are what we quote.
Verified before it publishes
Before an article is published, every figure in it is checked against the underlying data it was drawn from. The check is deterministic — a fixed rule, not another judgment call. Machine-generated and machine-revised article bodies are hard-blocked: they cannot go live with a number the data does not support. Later editorial corrections made by a reviewer are re-run through the same check, and any discrepancy is flagged for the reviewer to resolve.
Reviewed before it appears
A generated article is not published automatically. It waits for a person to review it and either approve it for publication or turn it down. Only approved articles appear on the site.
Corrections
When something in a published article needs to change — a figure to fix, a claim to sharpen — we revise the article in place and keep a record of the change rather than quietly overwriting it.
Spotted something that looks wrong? Please tell us — corrections from readers are welcome.
Not investment advice
Our research is for informational and educational purposes only. It is not investment advice, not a recommendation to buy or sell any security, and not a substitute for your own due diligence or a licensed professional. Figures can contain errors, and filings are sometimes restated.