Clay

Clay is a data-enrichment and prospecting platform built around a spreadsheet-style table where every row is a record (a person or company) and every column can run a lookup, an API call, or an AI prompt. You start from a list, a search, or one of its built-in sources, then layer enrichments on top: find emails, pull firmographics, scrape a website, score a fit, or write a personalised line with an LLM. It chains many data providers behind one interface and runs the whole thing on a schedule, so it sits between a lead database, an enrichment tool, and a lightweight automation engine.

Where it fits

Clay earns its place in outbound and revops work. The main use cases are building targeted prospect lists, enriching a CRM or a raw export with the fields that are actually missing, and generating the research that makes a cold message land instead of read like a mail merge. The "waterfall" enrichment (try provider A, fall back to B, then C) genuinely lifts match rates above any single data source, and the AI columns let you turn scraped context into a usable signal at scale.

It is genuinely for outbound teams, growth operators, and agencies who live in lists and care about data quality. It is less for someone who just needs a clean CRM and a handful of contacts, or for a team with no appetite to learn a tool that rewards tinkering. If your motion is inbound or product-led, most of Clay's power goes unused.

The honest take

The strength is range and composability: one canvas replaces a stack of point tools, and the credit model means you pay for enrichments that actually return data rather than a fixed seat per provider. The community and template library are strong, so you rarely start from a blank table.

The trade-offs are real. There is a learning curve: the table-plus-columns model is powerful but unfamiliar, and getting a workflow reliable takes iteration. Credit consumption is easy to underestimate, especially once AI columns and multi-step waterfalls run across large lists, so cost discipline matters. And the output is only as good as the upstream providers, so you still verify before you send. Clay has a free tier and paid plans; treat the pricing as something to model against your own volume rather than a flat number.

INTERVIEW EWOUD: What is your personal verdict on Clay in one or two sentences, and your star rating out of five?

INTERVIEW EWOUD: Is Clay in your own stack right now, and if so what do you actually use it for (and if not, why not)?

INTERVIEW EWOUD: What is the one line you would tell a peer who is deciding whether to adopt Clay?

B2B sales, growth and RevOps teams building enriched prospect lists from many data sources

Clay is a data enrichment and prospecting platform that combines a spreadsheet-style interface with waterfall enrichment across 100+ data providers and AI research agents. Teams upload or build lists of companies and contacts, enrich them with emails, phone numbers, firmographic and technographic data, and push results to a CRM or outreach tool. It is aimed at B2B sales, marketing, and RevOps teams that want to consolidate multiple enrichment sources and automate list-building without code.

This guide gets you from a blank Clay account to a working data engine: tables that pull in companies and people, enrich them automatically across dozens of sources, and hand off clean, scored records to the rest of your stack. It is for founders and growth operators who want to stop paying for ten separate enrichment tools and instead run one programmable layer that does sourcing, enrichment, and outreach prep in the same place. You do not need to be technical, though a willingness to think in spreadsheets and a little logic goes a long way.

Getting set up

The decisions that matter in Clay happen before you import a single row, so spend ten minutes on the foundations rather than diving straight into the data.

First, get clear on what a table represents. In Clay a table is a workspace where each row is one entity (a company or a person) and each column either holds data or runs an action. Decide upfront whether a given table is company-led (you start from accounts and find people inside them) or people-led (you start from a list of contacts). Mixing both in one table is where beginners create chaos, so keep them separate and link them deliberately.

Second, sort out your data sources. Clay's value comes from its enrichment providers, and most of them run on a credit system. You can connect your own keys for tools you already pay for so you spend your own credits instead of Clay's, which matters enormously for cost as you scale. Connect the integrations you genuinely use (your CRM, your email tool, LinkedIn-adjacent sources) early, because that is what lets Clay both pull data in and push it back out.

Third, understand credits before you build anything ambitious. Every enrichment call costs credits, and a table of a few thousand rows running several enrichments each can burn through an allowance fast. Set your plan to match the volume you actually intend to process, not the volume you imagine, and treat credits as the real currency of the tool.

How to actually use it

The core loop in Clay is import, enrich, condition, act, and you get the most value by doing those in order rather than bolting enrichments on at random.

Start by importing or sourcing your rows. You can paste a list, upload a CSV, pull from an integration, or use Clay's built-in sourcing to find companies and people that match a profile. Get the raw entities in first and resist the urge to enrich immediately.

Next, enrich in layers. Add a column that finds the company domain, then a column that pulls firmographics, then a column that finds the right person, then a column that finds their email. Each enrichment column reads from the columns before it, so the order is the logic. Build it like a pipeline where each step depends cleanly on the last.

Then add conditions so you only spend credits where it makes sense. Clay lets a column run only when a condition is met, so you can say "only look for an email if we found a valid LinkedIn profile" and avoid wasting calls on dead rows. This single habit is the difference between a tidy credit bill and a runaway one.

Finally, act on the clean rows. That might mean writing a personalised opening line with an AI column, scoring the row against your ideal-customer criteria, or pushing the finished record into your CRM or sequencing tool. The output of a good Clay table is not data, it is a ready-to-use record that something downstream can consume without a human cleaning it first.

Power moves

The AI column (often branded as Claygent or similar) is where Clay separates from ordinary enrichment tools. You can point it at a website or a prompt and have it read pages, extract a specific fact, and return it as structured data, which means you can answer questions no fixed provider sells, like "does this company run a careers page hiring for sales" or "what pricing model do they use".

Waterfalls are the second power move: instead of relying on one email-finder, you chain several so that if the first returns nothing, the next one tries, and the next. You pay for the result, not the attempts that failed, and your match rates climb sharply.

Use conditional runs aggressively to control spend, and use lookups to join tables so a company table and a people table stay in sync rather than drifting into two copies of the truth. And once a table works, save it as a template or reuse its structure, because the real compounding value is a library of proven tables you can rerun on new lists.

Where it fits your stack

Clay sits in the middle of a growth stack as the enrichment and preparation layer, not as the system of record. It takes a thin signal (a domain, a name, a list) and turns it into a rich, scored record, then hands that record off.

Upstream, it connects to your sourcing: lead lists, your CRM's existing contacts, sales-intelligence sources. Downstream, it pushes to your CRM, your outreach and sequencing tools, and your data warehouse or sheets. Its HTTP and webhook capabilities mean that if a native integration does not exist, you can still wire it to almost anything. The right mental model is a clean conveyor belt that feeds your CRM and outreach tools records they can trust, so those tools stop being where data goes to rot.

Pitfalls to avoid

The first trap is treating Clay as a CRM. It is brilliant at processing and preparing data, but it is not where your relationships and history should live. Push finished records out and keep your source of truth elsewhere.

The second is ignoring credits until the bill arrives. Run enrichments on a small test slice first, confirm the columns return what you expect, then scale to the full list. Building a huge table and running every column on every row before checking the logic is the classic expensive mistake.

The third is skipping conditions, which means paying to enrich rows that were never going to qualify. The fourth is building one sprawling table that does everything, when two or three focused tables are easier to reason about and cheaper to run. And the fifth is trusting enrichment output blindly: providers return wrong or stale data, so validate the fields that matter (deliverability on emails especially) before you act on them.

INTERVIEW EWOUD: Which enrichment sources and own-key integrations do you actually connect in Clay, and how do you keep credit spend under control at your volume?

INTERVIEW EWOUD: Walk me through the one Clay table or workflow you rely on most, step by step, including how it feeds the rest of your stack.

INTERVIEW EWOUD: What is your hard-won tip with Clay, the thing you wish someone had told you before you built your first serious table?