Wall Street Prompt · Brand

Where AI Actually Helps in Equity Research (Sorted by Task, Not Hype)

Instagram post adapted from a SPECIFIC captured source post — exact-source traceability, like our YouTube flow.

Instagram carousel

Validated — exact source linked@leadgenman

https://www.instagram.com/p/DWw_bgpjTX-/

Observed format: Carousel list post: a cover hook ("most AI tools are noise"), then category-by-category ranking of tools sorted by the job they do (coding, creativity, automation) rather than by hype, ending on a save-the-list payoff.

Observed hook: "I've tested 10s of AI tools over the past 2 years. Most are noise. These are the ones I actually use... Categorized by what you're trying to do, not by hype."

Adaptation: Kept the winning structure (anti-hype cover hook, job-to-be-done categories, "I actually use this" credibility, a deliver-the-list CTA) and changed everything else. Topic moved from generic AI productivity tools to AI inside an equity-research workflow. The source's tool tiers (coding, creativity, automation) became research tasks (sourcing, summarizing, comparing, question-drafting). Added a WSP-specific risk slide (WHERE IT BREAKS) that has no analog in the source, reframing the anti-hype stance as concrete failure modes (math from memory, price-target-sounding output, uncited claims). No source text, claims, tool names, or visuals reused. CTA changed from an implied save to a comment-keyword DM delivery. Added the educational-only, no buy/sell/hold disclaimer in WSP voice; removed all em dashes and hype language.

Cover

Cover slide

I tested AI across 50+ filings. Most of it is noise. Here is where it actually pulls weight in a research workflow.

Slides

One idea per slide

Slide 1

I tested AI across 50+ filings. Most of it is noise. Here is where it actually pulls weight in a research workflow. Sorted by task, not hype.

Slide 2

SOURCING. AI points you to the right section of a 10-K fast. Use it as a table of contents, not a reader. You still open the filing.

Slide 3

SUMMARIZING. Compress a 90-minute earnings call into themes. Then verify every claim against the transcript. A clean summary can still invent a number.

Slide 4

COMPARING. Build the comps skeleton: same metrics, same periods, same definitions. AI frames it. You confirm every figure pulls from the statements.

Slide 5

DRAFTING QUESTIONS. Generate the bear case and the questions you did not think to ask. This is where AI earns its seat.

Slide 6

WHERE IT BREAKS. Math from memory. Anything resembling a price target. Claims with no filing behind them. No source, no fact.

Slide 7

The rule: AI drafts, you verify. The workflow that wins is the one where every output traces back to a primary source. Comment WORKFLOW for the one-page version.

Caption

Paste under the post

Two years of running AI next to a real research process, and most of it is noise. The honest version is not a magic tool. It is knowing which task AI speeds up and which task it quietly breaks.

So here is the map by job to be done, not by hype:

SOURCING. AI is fast at pointing you to the right section of a 10-K or 10-Q. Treat it as a table of contents, not a reader. You still open the filing.

SUMMARIZING. Good for compressing a 90-minute earnings call into themes you then verify against the transcript. The risk is a clean summary that invents a number. Always check the source line.

COMPARING. Useful for setting up a comps frame: same metrics, same periods, same definitions across names. AI builds the skeleton. You confirm every figure pulls from the actual statements.

DRAFTING QUESTIONS. Strong for generating the bear case and the questions to bring into your own notes. It surfaces what you did not think to ask.

WHERE IT BREAKS. Math from memory, anything that sounds like a price target, and "trust me" claims with no filing behind them. If it cannot cite the source, it is a hypothesis, not a fact.

The pattern is simple. AI drafts, you verify. The workflow that wins is not the one with the most tools. It is the one where every output traces back to a primary source.

This is educational, not advice. No buy, sell, or hold here. Just process.

Want the one-page version of this workflow? Comment WORKFLOW and we will send it over.

Layout

Seven 1080x1350 slides. Slide 1 is the cover: large stacked headline, top two-thirds near-black, thin green underline rule under "noise," small WSP wordmark bottom-left, swipe cue bottom-right. Slides 2 to 6 share one template: bold green category label top-left (SOURCING, SUMMARIZING, etc.), one-to-two line body in white below, a faint slide counter (2/7) top-right, thin green left margin rule running full height. Slide 6 (WHERE IT BREAKS) flips accent to a muted amber rule to signal caution. Slide 7 is the payoff: centered rule statement, CTA line in green, WSP wordmark.

Design notes

WSP near-black background (#0B0F0E to #10140F). Primary text off-white #F2F4F1. Accent green #1FC77B used only on category labels, the underline rule, and the CTA keyword. Caution slide 6 uses a dimmed/low-opacity WSP-green rule (single-accent system â€” no second color). Montserrat throughout: ExtraBold for category labels and the cover headline, Medium for body, SemiBold caps for the WORKFLOW keyword. Generous left margin, left-aligned body, no center text except slides 1 and 7. Thin 2px green vertical rule on the left edge of body slides for a ledger feel. No icons, no stock imagery, no emojis on-image. High contrast, lots of negative space.

Why this works

The source wins on anti-hype authority plus utility: it sorts tools by job-to-be-done, claims real daily use, and promises a keep-able list. That structure maps cleanly onto WSP because the brand is already anti-hype and process-first. A "sorted by task, not hype" carousel lets WSP show operational credibility (10-Ks, earnings calls, comps) while staying educational and risk-aware. The category structure makes it skimmable and saveable, and the explicit "where it breaks" slide turns risk-awareness into a feature rather than a disclaimer. The comment-keyword CTA matches a proven list-delivery mechanic and builds a DM list.

CTA

Comment-keyword mechanic adapted from the source's implied save/list value: "Comment WORKFLOW and we will send the one-page version." Drives comments and a DM follow-up instead of a passive save.