What is Clay? How the clay.com GTM data platform works

What is Clay, and what does clay.com actually do?

Clay is a go-to-market (GTM) data platform that lets you build lists, enrich them and act on them inside one table. On clay.com, Clay describes itself as infrastructure to get any data, run agentic workflows and launch GTM plays. It is a different product from clay.earth, which is a personal relationship manager.

The core idea is simple. Every row is a company or a person, and every column is either data you imported or a step that fetches, calculates or sends something for that row. One column might find a work email, another might ask an AI model to summarise a company's pricing page, and a third might push the row into a lemlist campaign.

Clay's homepage says it is trusted by more than 500,000 GTM teams and names customers including Anthropic, OpenAI, Figma, Intercom and Vanta. Those are vendor figures, but they explain why Clay has become the default workbench for GTM engineering teams who build outbound systems instead of buying one contact database.

How do Clay tables, sources and enrichments fit together?

Every Clay table begins with a source, according to Clay's documentation. Sources include list builders such as Find Companies and Find People, CRM imports from HubSpot and Salesforce, webhooks and CSV files. Sourcing records and importing your own data do not consume Actions under the 2026 pricing model.

Once rows exist, you add enrichments. An enrichment is a column that calls a data provider or an AI model for each row, such as a LinkedIn profile lookup, a technographic check or a phone number search. Formulas and filters are free, so cleaning names, normalising domains and simple scoring cost nothing. The workflow most outbound teams run in Clay follows six steps:

1. Source companies that match your ideal customer profile.
2. Enrich those companies with firmographic, hiring or tech stack data.
3. Find people at the companies by job title.
4. Find and validate work emails through a waterfall.
5. Score each row or write a personalised line with AI.
6. Push qualified rows to a sequencer or CRM.

Each step is a column you can rerun, edit, switch off or restrict with a run condition. That structure is why Clay suits teams who want to test several targeting ideas in a week rather than commit to one list.

How does Clay waterfall enrichment work?

Clay waterfall enrichment runs several data providers in a set order for the same field, so a row one provider misses can be filled by the next. Clay's docs describe waterfalls as using multiple data providers in a predetermined sequence. For the Work Email waterfall, Clay states that you only pay credits for the provider that finds a match.

Coverage is the reason to bother. Clay's guide to finding work emails says the best single providers reach roughly 52 to 90% coverage depending on provider and market, while a three-provider waterfall reaches about 90 to 95%. These are Clay's own test figures rather than an independent benchmark, but the logic holds because each provider has different blind spots.

You can also put a free Infer Email step at the start, which builds an address from a pattern such as first.last@domain.com and stops the waterfall if validation passes. The full build, provider order and cost logic are in the Tentt guide to waterfall enrichment.

What is Claygent, Clay's AI research agent?

Claygent is Clay's AI agent for research and judgement calls a formula cannot make, such as whether a company sells to enterprises or what it launched last quarter. You build Claygents in Claygent Builder from a blank prompt, from a template, or by describing the task to Sculptor, Clay's conversational builder, which drafts the prompt, variables, output format and test cases.

Claygent browses the web when web search is switched on, or works only from data already in your table when it is off. Clay's docs say you can test up to 10 inputs at a time for free. Once saved, a Claygent is added to any table through an AI column, and edits in the builder update every table that uses it.

Claygent's cost depends on the model you pick. Clay's AI pricing docs list fixed prices for about 80% of models, for example 1 Data Credit for a Clay Helium web research task and 3 for Clay Argon. Token-heavy reasoning models are billed on actual token cost with 0% markup, and every AI prompt also counts as one Action.

Where does Clay send data: sequencers, CRMs and signals?

Clay sends enriched rows to the tools that act on them. Its documented integrations include sequencers such as lemlist, Instantly and Smartlead, where an Add Lead to Campaign action pushes each row into a campaign. For CRMs such as HubSpot, Clay can create, look up and update objects, and the Growth plan adds CRM auto-sync.

Clay also has its own email Sequencer. Clay's docs say the Sequencer uses one Action per lead added to a campaign, not per email sent. On the Growth plan you can buy sending domains and inboxes for the Sequencer inside Clay.

Signals run the process the other way: Clay watches for events and adds rows when they happen. Clay's signals docs list job changes, new hires, promotions, job postings, news and fundraising, topic intent and web intent, with web intent needing Growth or Enterprise. That makes Clay usable for signal-based outbound as well as static list campaigns.

How do Clay credits work after the 2026 pricing change?

Clay credits are split into two meters as of October 2026: Actions and Data Credits. Clay announced the change on 11 March 2026, separating data costs from platform work and cutting marketplace data prices by 50 to 90%. The table below summarises Clay's own documentation.

Clay's docs say a fully enriched record typically costs 6 to 20 Data Credits, depending on the data types, how many of your own API keys you connect and whether you run waterfalls. When a lookup returns nothing, Clay's pricing FAQ says no Data Credits or Actions are charged.

Customers on legacy Starter, Explorer and Pro plans can stay on them, but Clay's legacy plans doc says any plan change after 10 April 2026 moves you to the new plans. Plan prices and worked cost examples are in the Tentt Clay pricing guide.

Who uses Clay GTM workflows, and who should skip them?

Clay suits GTM engineers, RevOps and growth teams who want to combine many data sources and automate every step between finding an account and contacting it. The table model rewards people who think in workflows and are comfortable with formulas, conditional runs and API keys.

Clay is a poor fit for a founder who needs a few hundred contacts a month from one database. An all-in-one tool such as Apollo bundles search, emails and sequencing behind one login, and the learning curve is shorter. Clay's Free plan caps tables at 200 rows and excludes phone enrichment, so serious testing usually means the 14-day trial or the Launch plan.

The real trade-off is flexibility against simplicity. Clay gives you more data sources and more control than any single provider. You pay for that in setup time and in watching two credit meters instead of one.

What is Clay best for in your outbound stack?

Clay is best for teams who want to own their outbound data pipeline: source accounts, enrich them through waterfalls, research them with Claygent and push clean rows to a sequencer or CRM. It is the wrong purchase if all you need is a simple contact database.

If you are still asking what is Clay in practice, build a free table of 50 rows, run one waterfall and one AI column, and check both credit meters afterwards. Tentt builds Clay-based outbound systems for B2B teams, and the free GTM stack resources in the Tentt newsletter are a sensible next step if you want to go further.