Inside the Online Course

Learn how to research stocks with AI

This Course is for people who want to research U.S. public companies with AI using SEC filings, financial tables, and source checks — not stock picks or trading signals.

Online Course

AI for Stock Research Online Course

6 modules and 20 lessons. Each lesson has steps, a prompt to copy, and a hands-on exercise — plus a company research dossier you build across the Course.

$350

Buy the Online Course
Research notebook linked to a chart, magnifier, checklist, source stack, amber sticky note, mug, and a cyan check badge under a starry sky
Sample lesson

Lesson 1.3

Check an AI stock summary against the 10-K

Ask an AI to sum up a company's results and it can sound sure while mixing up the years. This prompt has the AI check a summary against the company's own figures, one claim at a time: it keeps what the filing supports, fixes wrong numbers and periods, and deletes advice.

You now have: A checked claim table and a corrected paragraph.

Copy the prompt

Help me check an AI draft against my sources.

Paste below: my sources, then the draft.

Do this:
1. Split the draft into claims.
2. For each claim, quote the source line that supports or contradicts it, or write NONE.
3. Check every number's period and unit against my sources.
4. Build a table: Draft claim | Source line (quoted) that supports or contradicts it, or NONE | Keep / Fix / Delete
- Put each claim on its own row; split a sentence that makes more than one claim.
5. Mark Delete for any sentence that recommends, rates, predicts or sets a price.
6. Rewrite the kept text using only supported claims, with periods and units.

Rules:
- Do not soften a deleted sentence into a new one.
- Research boundary: Use only the public source excerpts I supply. Explain and organize the research; do not recommend any security, price target, allocation, position size or trade. Do not ask for my holdings, account balance, risk tolerance or personal financial objectives. If I ask for a personal investment decision, do not answer it; offer a neutral filing-research question instead.
- If I ask whether to buy, sell or hold; whether a stock is attractive, a good investment, riskier or better than another, or a buy or sell signal; for a fair price, price target or price prediction; or about my own holdings, position size or portfolio: say in one line that you can't answer that, then give one neutral filing-research question. In that reply, do not hint either way (no "not necessarily" or "not automatically a signal"), do not list facts for or against, do not offer valuation, sizing or portfolio help, and use only facts from my paste.
- If I ask you to rewrite, soften or keep a sentence you marked Delete because it recommends, values or times a stock, treat it as a decision question under the rule above: one-line refusal, then one neutral filing-research question, nothing else.
- Keep = a source line supports it as written. Fix = a source line contradicts it or gives a different number or period; rewrite it with the source figures. Delete = no source line addresses it, or it is opinion, valuation, timing or a recommendation. Always Delete any sentence that recommends, rates, predicts or sets a price, even if it cites a pasted number.
- When the paste gives a cause, attribute it to management ("management said…"); don't state it as fact.

Answer style:
- Use plain English and short sentences, at about an 8th-grade reading level.
- Explain each finance term in one plain line the first time you use it.
- Give every number its unit and its date or period.
- Put any table first.
- Then add "What this means in plain English": 2–3 sentences of facts only. Give no advice and no verdict.
- Last, add "What to double-check": up to 3 items for me to check in the source.

Done when

  • Each Keep row quotes a source line
  • Period and units checked on every number
  • Advice-shaped sentences deleted
  • Two claims checked against the practice sources
  • Every row has Keep, Fix or Delete. If a row is blank or the table looks broken, ask the AI to redo the table with one claim per row.
More about this lesson
  • Then you check two lines yourself.
  • In the Course, this becomes your "Assumptions log 1."
  • Paste public filing excerpts and AI drafts only.
  • Try it now: copy the ready-to-run prompt into a new chat in the AI tool you already use, and send.
See the full lesson steps
  1. Open a new chat in the AI chat tool you already use.
  2. Copy the ready-to-run prompt and paste it into the chat.
  3. Add your sources and the draft under the prompt, then send it.
  4. Read the table first. Each claim should have a quoted source line or NONE, and Keep, Fix or Delete.
  5. Compare with the expected result: sales Keep (with "FY2025 vs FY2024"), net income Fix, cash flow Fix, margins Fix, "undervalued" Delete and "great year" Delete.
  6. Check the Keep row and one Fix row yourself. Find each quoted line in the sources. A Keep row you can't find becomes Fix.
Why this prompt works
  • Splitting the draft into claims forces every sentence through a source check before it stays.
  • Keep / Fix / Delete tags stop fluent wording from hiding a wrong period or an invented figure.
  • Delete (not soften) for advice-shaped lines keeps "undervalued" and "good entry point" out of your notes.
  • The research boundary tells the AI to stick to organizing the filing and not to recommend a security, a price target or a trade. If a reply drifts into advice, delete it.
  • "What to double-check" sends you back to the source for the rows that matter most.
Common mistakes and how to fix them
  • Keeping a figure because the sales line checked out. Fix: check every Keep row and at least one Fix row yourself.
  • Softening advice into a new sentence. Fix: Delete advice-shaped lines; do not rewrite them as milder opinions.
  • Mixing FY2024 and FY2025 numbers. Fix: give every number its period and unit from the source table.
  • Skipping the source quote. Fix: a Keep row without a quoted source line becomes Fix.

This is 1 of 20 lessons in the AI for Stock Research Online Course.

Buy the Online Course

Before and after

Before
Cedar Harbor had a great year: sales rose 7.4% to $2.48 billion and net income climbed to $78.7 million. Operating cash flow stayed above $150 million. Margins are improving as freight costs ease, and the stock looks undervalued — a good entry point before the holiday quarter.
After

Cedar Harbor Retail Co. is a made-up practice company, not a real SEC filing. AI replies vary, so check every figure against the source.

Draft claimSource lineKeep / Fix / Delete
Cedar Harbor had a great year.NONE. The source provides financial results but does not rate the year's performance.Delete
Sales rose 7.4%.“Net sales increased $170.0 million, or 7.4%, to $2,480.0 million in fiscal 2025.”Keep
Sales reached $2.48 billion.“Net sales increased $170.0 million, or 7.4%, to $2,480.0 million in fiscal 2025.”Keep
Net income climbed to $78.7 million.“Net income | 47.9 | 78.7” (PC-10K-FY25 §8, USD millions). FY2025 net income was $47.9 million, down from $78.7 million in FY2024.Fix
Operating cash flow stayed above $150 million.“Net cash provided by operating activities | 88.0 | 160.0” (PC-10K-FY25 §8, USD millions). FY2025 was $88.0 million, down from $160.0 million in FY2024.Fix
Margins are improving.“Gross margin decreased 100 basis points to 35.0%” (PC-10K-FY25 §7 MD&A).Fix
Freight costs are easing.“Gross margin decreased 100 basis points to 35.0%, primarily due to higher freight costs and promotional markdowns to reduce seasonal inventory.”Fix
The stock looks undervalued.NONE.Delete
The stock is a good entry point.NONE.Delete
The timing is before the holiday quarter.NONE.Delete

[… cut …]

Rewritten text using supported claims

[… cut …]

Net sales increased 7.4% to $2,480.0 million in FY2025 (52 weeks ended January 31, 2026).

Net income decreased to $47.9 million in FY2025, compared with $78.7 million in FY2024 (52 weeks ended February 1, 2025).

Net cash provided by operating activities decreased to $88.0 million in FY2025, compared with $160.0 million in FY2024.

Gross margin decreased 100 basis points to 35.0% in FY2025. Management said higher freight costs and promotional markdowns to reduce seasonal inventory were the primary reasons.

What this means in plain English

Sales increased during FY2025, but net income and operating cash flow decreased compared with FY2024. Gross margin also declined by 100 basis points. Management identified higher freight costs and promotional markdowns as the primary reasons for that decline.

What to double-check

1. Net income: Confirm the FY2025 figure of $47.9 million against the consolidated statements of operations.
2. Operating cash flow: Confirm the FY2025 figure of $88.0 million against the consolidated statements of cash flows.
3. Gross margin: Confirm the 35.0% figure and the 100-basis-point decline in the FY2025 MD&A.

Research boundary: This review uses only your supplied practice-company excerpts. The company and figures are fictional course-exercise material.

ChatGPT is a trademark of OpenAI. Siqwes is not affiliated with or endorsed by OpenAI.

What you'll leave with

  • A company research dossier you assemble with sources, periods, units and explicit evidence gaps.
  • Copyable financial tables: a cash-flow bridge, dated peer comparison and separate assumptions card (not a binary workbook).
  • Earnings and headline checks, a contradiction log, weekly research tasks and a separate unseen transfer exercise.

What's in the Course

  • Module 1: Research briefs that hold up

    3 lessons · 75–95 min

  • Module 2: Filing workflows on EDGAR

    3 lessons · 60–75 min

  • Module 3: Financial statements and comparison discipline

    4 lessons · 85–105 min

  • Module 4: Earnings, news, and catalysts

    4 lessons · 80–100 min

  • Module 5: Discovery, technical methods, and IPOs

    3 lessons · 60–75 min

  • Module 6: Run your research desk

    3 lessons · 65–80 min

Each lesson separates a small beginner core from required longer company-project work. These are estimates, not measured learner times. Each lesson has conceptual teaching, a worked input and expected answer, a prompt, an exercise and acceptance criteria. Every module ends with a playbook page.

How to get started

After payment, sign in on the members page and start learning right away.

Buy the Online Course

Common questions

Do I need a paid AI plan?

The core exercises are written for a no-cost chat plan and public SEC EDGAR filings; no paid terminal is part of the required path. Free tiers can limit usage, uploads and conversation length, and availability varies by provider and region. Use short excerpts and check the provider’s current plan and region pages.

How do I get in?

Sign in at members.siqwes.com with the email you used to pay. Your lessons are under My courses.

What stock-research work will I build?

A company research dossier with filing extracts, an income snapshot, a cash-flow bridge, a peer table, earnings checks, a contradiction log, and a weekly research checklist.

Does the Course give stock picks or trading signals?

No. It teaches research methods using public information. It does not recommend securities, price targets, allocations, position sizes, or trades.

How will I compare financial figures fairly?

You will record units, actual fiscal start and end dates, and definitions; convert thousands to millions; and note window overlap and business differences. The same fiscal-year label alone is not enough.

Do I need the free stock-research series first?

No. Module 1 starts from a complete research-brief template. The five-day series is an optional way to try the method first.

Free email series

Not convinced? Try it for free

Start with the free 5-day AI for Stock Research educational series.

View the free 5-day series

Educational only. Siqwes teaches research methods. Nothing here is financial, investment, tax or trading advice, or a recommendation to buy, sell or hold any security. Investing involves risk, including loss of principal. AI tools can be wrong, so verify everything against the primary source, such as the company's filings on SEC EDGAR, and use AI tools at your own risk. For decisions about your own money, consult a qualified, licensed professional. See our Terms, §8.