From filings to a published report

  1. Company filings and financial data

    Research starts from reported numbers and company disclosures, with the dates they became available.

  2. Deterministic screening

    Fixed, rule-based checks look for changes in growth, margins and cash flow. The same data always gives the same result; no AI model takes part in this step.

  3. Deeper research and evidence checks

    Companies that pass the screen are studied in more depth. AI tools help read filings and draft notes; every claim must trace back to a source.

  4. Editorial review

    A draft is checked against the original filings for figures, dates and periods, and for risks and contrary evidence. Model research notes are published as they are, clearly labelled as unreviewed AI interpretation; reports are reviewed.

  5. Publication

    Reports are published with their data dates and sources, and updated or corrected when the evidence changes.

Screening and interpretation are kept separate

Which companies are looked at is decided by the rule-based screen. AI models assist with reading and explaining the evidence afterwards; their output is interpretation, not a score, and it can be wrong.

Candidates are not recommendations

A company appearing in the research is a subject for study, not an instruction to buy or sell. Research here is general information, not personalised financial advice.

The detail behind each step is in the methodology, and the rules every report follows are in our editorial principles.