In detail / 02 · RESEARCH
One question, taken to the evidence.
Ask with material and context: read the filing, pull the multi-year data, check the original announcement — and write conclusions with their basis and verification stated. Verified, inferred and doubtful claims stay separate; guesses never pose as facts.
- Analyze filingsStart from one quarter's numbers: what is operating improvement, what is one-off, and what a single quarter cannot answer.
- Pull dataRevenue, margins, capex and more, listed year by year — with differing calculation bases reconciled and the gaps explained.
- Track holdingsAgainst the original SEC 13F filings: adds and trims are judged on share counts, and value moving the other way is flagged as price.
- Check the original filingAnnouncements and prospectuses verified against the original document, with links so you can recompute yourself.
- Sessions that persistMaterial and dialogue stay in place; the next research session continues where the last one ended, with its basis one look away.
Scene 01 / Analyze filings
“How did Apple do on revenue and profit last quarter — and what can one quarter not answer?”
Facts pulled from the filing
Apple FY2026 Q3 (ended 2026-06-27): revenue $109.417B, net income $29.789B. Next come segment breakdown, cash flow and the year-ago quarter — two numbers alone cannot price a business.
FMP · AAPL income-statement · USD, quarterly; filed 2026-07-31
Historical research excerpt, not live data. Only part of the answer is shown; read it alongside the original material.
Another complete public research example: do configurations prove execution?
How to research a company
Start with a question, examine the material, form a view, and identify what still needs checking.
- Your questionWhich company, which question, and which period?
- Original materialCheck document links, publication and retrieval dates, and the relevant passages.
- Your viewSeparate conclusions supported by evidence from assumptions and evidence against them.
- Further checksWhat material is missing? Which definitions differ? What should you check next?
This illustrates an answer’s structure, not a full research report or a current valuation.
What research does not do
- No buy/sell signals and no promised returns — research supports your judgment, it does not replace it.
- Evidence carries its grade: verified, inferred and doubtful stay separate; guesses never pose as facts.
- Technical progress alone is not an investment opportunity — price, uncertainty and the business still require investigation.
Before you start
- What it is
- Reading and research tools for investing in AI.
- Who it is for
- Long-term investors who want to understand AI businesses and industry changes, and form views backed by evidence.
- What you can do
- Read public material, explore the AI industry map, AI 100 and investment tools. Sign in for personal research and long-term watches.
- How to use it
- Start with a question and relevant material, then review the answer and its sources. For a watch, set the subject and frequency, review the plan, and confirm.
- Sign in
- Public reading and resources need no account. Saving, following, personal research and long-term watches require Google sign-in.
- Cost
- The public parts are currently free. For signed-in research and watches, check the costs and usage limits shown when you start.
- Usage limits
- New accounts include trial credits. Check the limits and conditions shown when you start research or a watch.
- Sources and dates
- Materials keep source links and dates. Check whether a date refers to the edition, publication or verification. Research answers explain their sources and missing material. Financial snapshots and historical excerpts are not live quotes.
- Keep in mind
- No buy/sell signals, promised returns or managed funds. Even with a source link, check the original material and any unresolved questions.