Methodology

Each company is scored 1–5 (an Unknown score is never averaged in) on seven dimensions (higher is always better / more durable). Missing or private data is recorded as Unknown, never estimated, and lowers confidence.

• Real customers & revenue quality: Are there named paying customers, recurring revenue, and retention evidence?
• Unit economics: Do gross margins survive inference/compute costs? Is pricing durable?
• Moat: Proprietary data, distribution, switching costs, or tech that is hard to copy?
• Platform risk: Could a model provider or incumbent ship this as a feature? (higher score = lower risk)
• Team & execution: Shipping cadence, leadership track record, stability.
• Funding vs. traction: Is capital raised justified by verifiable traction?
• Hype signals: Demo-driven marketing, vague metrics, valuation outrunning evidence? (higher score = fewer red flags)

Verdicts

• Buy: durable — safe to build on / commit to.
• Hold: promising, but key evidence is missing; pilot with short contracts.
• Hype: attention outruns verifiable traction or the product is at high platform risk.
• Not yet rated: research pending.

Confidence

Low: fewer than 4 dimensions have cited evidence. Medium: 4–6. High: all 7, with at least one primary source (filings, company disclosures, named customer case studies).

Sourcing policy

Every factual claim links to a source. We do not fabricate revenue, user counts, valuations, or quotes. Reported figures are attributed to who reported them (e.g. “per Reuters”). Not investment advice.

Paid work

A Stack Review or a vendor-selection report never changes a public grade. Vendors cannot pay to alter, delay, or remove a verdict. Affiliate links are not used today; if one is added, the page will say so. Details are on the pricing disclosure.