AI for Equity Research: From Retail Investor to Junior Analyst
Filter earnings calls and filings with AI, read margins and DuPont, value with P/E and P/B bands, pitch on one page. 8 lessons, first 2 lessons free.
On 15 October 2026 TSMC reported third-quarter revenue of NT$1,494.24 billion. Anyone who had added up its monthly revenue reports for July, August and September already had that number: Taiwan’s listed companies report the previous month’s revenue by the 10th. The information an analyst needs is public. What separates a junior analyst from a reader is how fast they filter it, how carefully they check it, and whether they can say what the share price already expects.
This course follows TSMC through one results season. You filter the release, the call and the filing with ChatGPT, Claude or Gemini and trace every number to its page. You compare results with a dated consensus, read the gross-margin trend, DuPont ROE and free cash flow from a data platform such as Statementdog, place the valuation on P/E and P/B bands and a comps table, and write bear, base and bull cases with kill criteria. Then you follow the signal from Taipei to Tokyo Electron, Advantest, SK hynix and Samsung, and finish with a one-page pitch built on a copy-ready template.
The AI reads, sorts, drafts and argues back. It does not get the last word on a number, and it never writes your thesis. This is an educational course in research method, not investment advice. Built at the request of a Pro member. Lessons 1 and 2 are free.
What You'll Learn
- Apply an AI-assisted results-day triage to an earnings release, a call transcript and a filing, with every number traced to its page
- Analyze what the share price already assumes by comparing results and guidance with a dated consensus, and state a variant perception with a dated catalyst
- Analyze a company's gross-margin trend, DuPont ROE and free cash flow from a visual data platform, checking the base of every figure
- Evaluate P/E and P/B bands and a comps table for a cyclical company, normalizing earnings and aligning fiscal years, currencies and accounting standards
- Create bear, base and bull scenarios on two or three stated drivers, with kill criteria and a risk/reward ratio
- Create a one-page investment pitch that separates fact from opinion and ties Taiwanese, Japanese and Korean supply-chain signals to the thesis
After This Course, You Can
What You'll Build
Course Syllabus
Prerequisites
- You already know what a stock, an income statement and a P/E ratio are: the course skips the basics
- Access to ChatGPT, Claude or Gemini (Lesson 1 lists each one's upload limits by plan) and a browser for filings and data sites
Who Is This For?
- Serious retail investors who read annual reports and want to work the way a buy-side or sell-side analyst does
- Aspiring equity analysts preparing for stock-pitch interviews or a research challenge
- Investors in Taiwanese, Japanese and Korean technology stocks who want to read the supply chain across the three markets
- Finance professionals who use AI at work and want a disciplined way to check what it tells them
Frequently Asked Questions
Is this investment advice?
No. The course teaches a research method. Its running case, TSMC, uses the company's public results for the third quarter of 2026 to show each step, the valuation inputs in the worked examples are labelled illustrative, and no lesson tells you to buy or sell anything.
Can I trust ChatGPT, Claude or Gemini with the numbers?
Not without checking. In the FinanceBench study (2023), GPT-4-Turbo with a shared document store answered 19% of 150 questions about company filings correctly and refused or got the rest wrong. The AI is fast at finding, sorting and drafting, so the course has you trace every figure to its page in the filing before it goes into your notes.
Which markets and filings does it cover?
US filings (10-K, 10-Q and the 8-K earnings release), Taiwan's monthly revenue reports and quarterly statements on MOPS, Japan's kessan tanshin on TDnet and securities reports on EDINET, and Korea's DART. Lesson 7 follows how a signal in one of these markets shows up in the others.
Do I need Excel or a paid data subscription?
No. There is no multi-year model to build. You work from filings, an AI assistant and a visual data site, and you check what it shows against the filing. On 13 November 2026, Statementdog's free sign-up showed 20 years of history and Koyfin had a free plan with two years of financials.
Is this course free?
Lessons 1 and 2 are free with a free account. The rest of the course and the certificate come with Pro.
Will I get a certificate?
Yes, with Pro: finish every lesson and score at least 70% on the quizzes for a verifiable certificate.
How long does it take?
Eight lessons; the course page shows the total time. Lesson 2 gives you a results-day digest and a number trace table you can reuse every quarter.
