Most AI research fails for one boring reason. It mixes three things that need three different standards of proof.
Fact: what the source literally says. Inference: what you concluded by connecting facts. Opinion: what you think it means or what it's worth.
Welded into one confident paragraph, they read true. That is exactly how a memo misleads you. The sentence sounds settled, and you stop asking which half is actually verifiable.
The fix is a labeling pass, AI labels and you judge:
1. FACT: stated in the filing, transcript, or comps, with a page or timestamp. No source, not a fact. 2. INFERENCE: a conclusion built from facts, with the assumption written down next to it. This is where most errors hide. 3. OPINION: a view on quality, direction, or value. It carries no source and AI does not get to pass it off as truth.
Then read in that order. Verify the facts. Stress-test each inference against its named assumption. Own the opinions or cut them.
AI builds the labeled draft. You decide what survives. A line with no label does not get used.
Save this and run it before you trust your next memo.
Educational content only. Not investment advice, and not a recommendation to buy, sell, or hold any security. Wall Street Prompt. Always verify against the primary source.