Clay

What it is

Clay looks like a spreadsheet and works like a research pipeline. Each row is a company or a person, and each column can call a data provider, run a calculation or ask an AI agent a question. Its defining feature is the waterfall: instead of trusting one data vendor for an email address, it tries providers in sequence across more than 150 sources until one returns a result.

The second half is Claygent, an AI research agent that visits sources and answers a question per row, such as whether a company runs a particular technology or recently opened a new location. That turns research that would have been an afternoon of manual checking into a column.

Who it's for (and who it isn't)

It fits sales and growth teams whose targeting depends on signals rather than firmographics: companies hiring for a specific role, businesses using a particular platform, accounts that just did something worth reacting to. If your outbound is only as good as your list, this is where lists get better.

It is a weaker fit for teams who have not yet defined who they sell to. Clay makes an unclear targeting hypothesis expensive rather than clearer, since every enrichment consumes credits. It is also not a sending tool: you build the list here and send it from Instantly or lemlist.

Pricing

Clay is freemium and metered on actions and data credits. Free gives 500 actions and 100 data credits per month, limited to 200 rows per table, and excludes phone enrichment, signals, CRM integrations and webhooks. Launch starts at $167 per month billed annually, or $185 monthly, with 15,000 actions and 2,500 data credits. Growth starts at $446 per month annually, or $495 monthly, with 40,000 actions and 6,000 credits. The lower "starts at" figure on Clay's pricing cards covers only the actions; the data credits come on top. Enterprise is custom.

The gap between annual and monthly pricing here is unusually wide, so a monthly commitment costs roughly three times the annual rate. And because actions are consumed by experimentation, budget for the tests that fail as well as the runs that work.

Set it up

Official guide: Get Started in Clay

1. Create a Clay account on the Free plan. It gives you 100 data credits per month (credits are what Clay spends when it buys data for you, such as an email address) and up to 200 rows per table (one row is one person). Keep your first list to at most 25 people, because each email address Clay finds uses credits. Your account now starts with 100 credits.
2. On the home screen, click "Find people". Clay has an AI helper called Sculptor: type in plain English who you want to reach, for example "marketing managers at software companies in London". You now see a list of filters that Sculptor filled in for you.
3. Read each filter Sculptor added (industry, seniority, location) and fix any that look wrong. If an exact job title finds too few people, switch it to "job titles similar to". Narrow the filters until the number of people found is 25 or fewer. The search now shows your short list.
4. Click "Continue", then save the results to a new workbook and table. A workbook is a file that holds your tables, and a table is the spreadsheet-style list of people. You now see your people as rows in a table.
5. Open the "Tools" menu in the top-right corner of the table and add the "Work Email" enrichment (a lookup that adds information to each row). It is a waterfall: Clay asks one data provider after another until one returns a valid email address, and you spend credits only for addresses it finds. Clay now fills a new column with email addresses.
6. Click "Use AI" to add a column where Clay's AI looks something up for each person, such as their recent public writing. Read every result yourself, because the AI can be wrong. The new column now shows one AI answer per person.
7. Open the "Tools" menu again, then the "Export" panel, and click "Download CSV". A CSV is a plain spreadsheet file that Excel or Google Sheets opens. Your browser now saves it to your downloads folder.

When this works, you have a downloaded spreadsheet of up to 25 people with work email addresses and one AI-researched column that you checked yourself.

Strengths and limitations

The strength is the waterfall plus research in one place. Coverage from many providers beats any single vendor, and paying per successful lookup rather than per subscription is the right shape for spiky work.

The limitations are cost control and verification. Credits disappear during experimentation, and it is easy to spend a lot learning the tool. AI research returns confident answers that are sometimes wrong, so a sample check is not optional. And the free plan's 200-row cap makes it a demonstration rather than a working tier for real campaigns.