AI research disclaimer

The agents are useful. They are not oracles. Here is precisely what that means.

What the output is

Agent findings are automated analysis of filings and market data. They are generated by software, reviewed by no human before you see them, and can contain errors — including errors of parsing, where a figure was read from the wrong line of a document.

How to read a score

• A research score is a weighted summary of the analyses that had enough data to run. It is not a prediction of price.
• When coverage is too thin, the score is withheld and labelled RESEARCH SCORE: UNAVAILABLE with the coverage shown. That is deliberate — no score is better than a confident-looking wrong one.
• A partial score is labelled partial. Treat it as a partial view.

Scenarios are not forecasts

Projected scenarios are AI-generated illustrations of what would follow from stated assumptions. Change the assumptions and the scenario changes. They are not predictions, and nothing about them is guaranteed.

Missing and conflicting data

Where a figure is unavailable the platform says DATA NOT AVAILABLE instead of estimating. Where sources disagree it shows DATA CONFLICT DETECTED and presents both. An absent number is information; a fabricated one is not.

Your decision

Use the research to understand a company and to know which questions to ask. The decision, and the responsibility for it, remains yours.