Research a hypothetical company responsibly

Research begins with a question

Use a fictional company: Northline Bikes, which makes commuter bicycles.

Start with questions instead of a verdict:

• How does the company make money?
• Who are its customers and competitors?
• Are sales, profit, and cash moving in the same direction?
• What does it owe, and when?
• What could change demand or costs?
• Which facts would challenge the optimistic story?

Build an evidence ladder

For a real public company, prefer original regulatory filings and audited statements over a social post or AI summary. Canadian public filings are generally found through SEDAR+; U.S. public-company filings are found through EDGAR.[1][2]

A useful order is:

1. regulatory filings and official disclosures;
2. regulator or government data;
3. independent, established reporting;
4. company marketing;
5. commentary, social posts, and AI summaries.

Lower on the ladder does not always mean false. It means you should demand more verification.

Give AI a narrow job

AI can help create a glossary, turn a table into questions, or compare two versions of a statement. Do not upload private documents or rely on it to quote a filing accurately.

Example prompt:

Here is an invented income statement. Explain the difference between revenue and profit, identify three questions the table cannot answer, and do not evaluate the company as an investment.

Keep a contradiction column

Create three columns: supports the story, challenges the story, and still unknown. A good brief includes all three.

Finish with uncertainty

Do not end with “buy” or “sell.” End with:

• what the evidence currently shows;
• what remains uncertain;
• what new information could change the view;
• which facts were independently verified.

That is responsible research—even when it produces no exciting conclusion.