Clay

Clay is a spreadsheet-style table where each row is a lead, company, or account and each column can call a data provider, run an AI research step, or trigger an outbound action (email, webhook, CRM write). Its core mechanic is 'waterfall' enrichment: instead of picking one data vendor, a Clay column can try multiple providers in sequence until one returns a match, which independent reviewers on Trustpilot specifically called out as its strongest feature ('the customizability... nothing else on the market compares for this level of enrichment logic', per a July 2026 review). Claygent, its AI research agent, is built on OpenAI models — confirmed independently by OpenAI's own customer case study, not just Clay's marketing.

Verified directly on clay.com/pricing (17 Aug 2026): Free ($0/mo, 500 Actions + 100 Data Credits/mo, up to 200 rows/table, unlimited seats), Launch ($167/mo, 15,000 Actions + 2,500 Data Credits/mo, up to 50,000 rows/table), Growth ($446/mo, 40,000 Actions + 6,000 Data Credits/mo, marked 'Recommended'), and Enterprise (custom, 100,000+ Actions and Data Credits typical, annual commitment). Every plan gets unlimited seats and tables, which is unusual for this category — you don't pay per user. The catch is that spend is metered on two separate currencies: 'Actions' (each enrichment or automation step) and 'Data Credits' (the cost of the data itself, from $0.05 each), and a buyer has to model both together to predict a real bill. This is a legitimate design for usage that varies wildly by workflow, but it also means the sticker price understates cost for anyone running complex, multi-step columns — a pattern that shows up repeatedly in independent reviews (below).

Clay's Trustpilot profile is rated 'Poor' — 2.4/5 from 14 reviews as of 17 Aug 2026 (small sample; treat accordingly). Reading the reviews directly rather than a star average: the positive review praises the enrichment logic and flags credit-usage vigilance as the trade-off ('you have to keep an eye on your credit usage so you don't burn through them while experimenting'). The negative reviews cluster around three specific, recurring complaints rather than one-off gripes: credits consumed by retries or platform behavior the reviewer didn't expect ('stealing credit usage by running stuff that was never intended to run'), data loss when editing a table's column filters, and slow or unhelpful support at higher spend levels (one reviewer cited $350/month with a support reply to 'ask your LLM for help debugging'). G2 also lists Clay reviews, but G2's review pages blocked direct verification (CAPTCHA) and search-engine snippets returned inconsistent numbers across what appear to be duplicate G2 listings for the product — we can't currently verify a specific G2 rating and aren't citing one.

Clay markets itself around the 'GTM engineer' role it says it popularized in 2023 — someone technical enough to build and maintain enrichment/automation logic, not a rep who logs in to send emails. That framing is corroborated by the negative reviews above, which mostly come from users who ran into complexity or cost surprises rather than missing features. If your team doesn't have (or want) someone in that operator role, budget for the ramp-up or expect the learning-curve friction that recurs across independent reviews.

Clay (Clay Labs Inc.) raised a $100M Series C led by CapitalG (Alphabet's growth fund) in August 2025, valuing the company at $3.1B, with Sequoia Capital, Meritech, First Round, BoxGroup, Boldstart and Sapphire Ventures participating — reported by TechCrunch and confirmed via Businesswire's press release. TechCrunch reported (Aug 2025) more than 10,000 customers including OpenAI, Anthropic, Canva, Intercom, Cursor and Rippling; these are company-reported figures relayed by the outlet, not independently audited. A later employee tender offer reportedly valued the company near $5B in January 2026, per secondary sources (Tracxn, Sacra) we could not independently confirm against a primary filing — treat that later figure as unverified.

Clay publishes a real trust center (trust.clay.com, Vanta-powered) listing SOC 2 Type 2 and Type 1 reports, ISO 27001:2022 and ISO 42001:2023 (AI management system) certificates, and GDPR/CCPA compliance status, plus a public subprocessor list (AWS, Anthropic, Amplitude, Braintrust and others) and named security controls across infrastructure, product, and AI-specific categories. A DPA is available on request via [email protected]. This is a meaningfully more complete disclosure than most tools in this category publish without a sales call.