Watch Equity Research Prompt Pack actually work

A quick note before the demo: this pack works on a company you already have in mind — you feed it a 10-K, it structures the teardown. It won't scan the market and hand you a list of fresh strategy ideas from scratch. So I'll show you what it actually does with a real filing, and then tell you honestly where "new strategies" fits.

The prompt (one of 45 — the thesis synthesis step):

"Using the 10-K sections I've pasted (Business, Risk Factors, MD&A, Financial Statements), build a one-page investment thesis. Output: (1) the core bull case in 2 sentences, (2) three specific catalysts with rough timing, (3) three material risks ranked by severity, (4) a price target with the multiple and the assumption it rests on. Flag any claim you can't support from the filing."

Running it live on Datadog (DDOG), FY2023 10-K:

Bull case — Datadog is a land-and-expand observability platform where existing customers keep buying more modules; 83% of customers now use 2+ products (up from 81%), and net revenue retention stayed above 110%, meaning the installed base grows spend without new logos.

Catalysts:
▸ Next 2–3 quarters — AI-native customers scaling infrastructure monitoring; management cited this cohort as a growing % of new bookings.
▸ 6–12 months — Cloud spend "optimization" that pressured 2023 growth appears to be normalizing; a reacceleration in usage-based revenue would re-rate the stock.
▸ 12–18 months — Security products (Cloud SIEM, CSM) maturing into a second growth engine beyond core monitoring.

Risks, ranked:
1. Valuation — trades at a high revenue multiple; any growth miss compresses it hard.
2. Concentration in usage-based billing — revenue falls fast when customers cut cloud consumption (proven in 2023).
3. Competition from Microsoft/AWS bundling native tools that undercut on price.

Price target — $135, applying ~14x forward revenue (a discount to its 2021 peak but a premium to slower SaaS peers), resting on the assumption that NRR holds above 110% and usage reaccelerates in H2. Unsupported flag: the pack cannot verify forward estimates — the multiple is a framework input you set, not a number pulled from the filing.

Honest assessment: if by "new strategies" you mean a screening system that generates candidates, this isn't that. What it is: once you have a name, it forces disciplined, repeatable analysis — the same 45 steps every time, so your process becomes the edge rather than your gut. Pair it with a separate screener for idea generation, and this handles everything downstream.