Wall Street PromptBrand
All briefs
Wall Street Prompt · Brand

The Prompt That Separates Fact, Inference and Opinion

Carousel — cover + one idea per slide + caption. This is a post, not a video script.

Carousel
Instagram carousel (+ LinkedIn document)Instagram carousel8 slides
References & validationReference missing

No specific creator reference resolved. Add an observed post before production.

Reference creator: Research methodology creators

No reference links resolved.

Slides

One idea per slide

Cover
FACT. INFERENCE. OPINION.

The prompt that keeps your AI research from lying to you.

Most memos mix all three.
The good ones label them.

AI labels. You judge.
(swipe)
Slide 1
FACT. INFERENCE. OPINION.

The one prompt that keeps an AI memo honest.

Most research fails because it blends three things into one confident sentence.

AI labels. You judge.
(swipe)
Slide 2
The common mistake:

You ask AI to "analyze the filing" and it hands back smooth prose.

The failure mode:
Fact, inference, and opinion are now welded together.

The memo reads true.
You can no longer tell which parts are.
Slide 3
Define the three before you trust any of them.

-> FACT: stated in the source. Page or timestamp.
-> INFERENCE: a conclusion you built from facts.
-> OPINION: a view on quality, direction, or worth.

Three different burdens of proof. One sentence hides that.
Slide 4
The prompt: make AI label every line.

-> [FACT] only if it's in the filing, transcript, or comps
-> [INFERENCE] tag it, then write the assumption next to it
-> [OPINION] tag it, and admit it carries no source

No label = not usable.
Slide 5
1. FACTS

-> Each one cites a page, a line, or a timestamp
-> If AI can't point to the source, it's not a fact yet
-> Numbers get a unit and a period
-> You verify these against the primary document, not the summary
Slide 6
2. INFERENCES

-> This is where most errors live
-> Every inference names the assumption it stands on
-> Change the assumption, does the conclusion still hold?
-> A weak assumption is a follow-up, not a finding
Slide 7
3. OPINIONS

-> "Strong moat." "Cheap." "Management is solid." All opinion
-> AI is not allowed to pass these off as facts
-> They are inputs to your judgment, not conclusions
-> You own every opinion in the memo. Sign your name to it
Slide 8
Then read in order.

-> Verify the facts
-> Stress-test the inferences against their assumptions
-> Own the opinions, or cut them

AI built the labeled draft. You decided what survived.

Save this. Run it before your next memo.
Educational only. Not advice.

Caption

Paste under the carousel

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.

Design notes

WSP system across all slides: near-black background, off-white body, one green accent reserved for the three labels (FACT / INFERENCE / OPINION), the bracket tags, and slide numbers. Cover uses the heavy condensed headline at max weight with the three words stacked and tight. Body slides follow one fixed grid: bold numbered/labeled header at top in accent, then a left-aligned arrow list (->), high text density like a lecture slide, never a minimalist tip card. Keep type sizes consistent slide to slide so it reads as a deck. Bracket tags on slide 4 set in a mono-styled face to look like a prompt spec. "No label = not usable." and the closing rules get extra weight and breathing room. Persistent small @WallStreetPrompt handle in a footer corner; final slide carries the "Educational only. Not advice." tag. One accent color only, no gradients, simple -> arrows throughout.

CTA

Save this and run it before you trust your next memo.

Wall Street Prompt
Wall Street Prompt — internal