Clay for B2B Outbound: Enrichment Workflows That Don't Burn Your List
Use Clay to enrich only leads that fit your ICP: filter first, waterfall emails with validation, run on empty fields only, and cap volume before you send.
7 min read

Contents
To use Clay for outbound without burning your list, enrich only the accounts and people that fit your ideal customer profile, find emails with a validated waterfall, run each enrichment only on rows that still need it, and send to a small, checked segment first. Enrichment decides who you contact; deliverability decides whether you can keep contacting anyone.
Picture a five-person B2B software company that imports 20,000 contacts from a list builder into Clay on a Monday. Every column runs on every row. By Wednesday the table has emails for most of them, the sequencer is loaded, and the first send goes out to everyone at once. Half the list was never a fit, a share of the addresses were catch-all guesses, and the complaint rate climbs within days. The data was enriched; the list was burned.
What is Clay, and what does an enrichment workflow do?
Clay is a spreadsheet-style platform that pulls data about companies and people from many third-party providers, runs AI research on each row and sends the results to other tools. An enrichment workflow is the ordered set of columns that turns a raw list into a short, accurate list worth contacting.
Enrichment means adding missing facts to a record: company size, industry, the person's role, a work email, a recent hiring signal. In Clay, each step is a column in a table, and each column can call a data provider, an AI model or another tool.
A sound workflow runs in this order:
Define the fit. Write down the company and role criteria that make a lead worth contacting.
Enrich the company first. One lookup per account removes the non-fits before you pay for lookups on each person there.
Filter. Drop every account that fails the fit criteria.
Find the people. Only at accounts that passed, find the roles you sell to.
Find and validate emails. Use a waterfall with validation switched on.
Research and write. Add the one or two facts that make a message relevant.
Push a small batch. Send a reviewed segment to your sequencer or CRM.
How does Clay pricing work in 2026?
Clay bills two things: Actions, which measure the work the platform does, and Data Credits, which pay for data and AI bought from providers in Clay's marketplace. Every enrichment uses an Action; it uses Data Credits too when Clay supplies the provider.
Clay's Actions and Data Credits guide says Actions "measure the orchestration you do in Clay: enriching data, running AI research, and sending data to other tools," while Data Credits "are used to buy data or AI from 3rd party vendors in Clay's data marketplace—costs vary by data type."
What this means for cost
Your own API keys. Connecting your own provider keys "will eliminate the Data Credit cost," but "an Action will still be consumed."
Empty results. Clay's FAQ on unsuccessful searches says you are charged Data Credits only "when Clay successfully returns data," and the same applies to Actions.
Validation is charged. The same FAQ says a validation check that returns "either a valid or invalid result" uses Data Credits and an Action, unless you use your own keys or a validator Clay offers free.
Actions don't roll over. The guide says "Unused Actions expire at the end of each billing cycle." Data Credits on monthly plans roll over, capped at twice the monthly amount.
Top-ups cost more. One-time Data Credit top-ups carry "a 30% premium."
Plan prices and allowances change, and on the day we checked, Clay's pricing page showed different starting prices in its plan cards and its FAQ. Read the current page before you budget, and size a plan from a test run on a few hundred rows rather than from a list price.
How do you set up an email waterfall in Clay?
A waterfall is a column that asks several email providers in turn and stops at the first valid result. Set it up with validation on, a conservative risk level and an inferred-email step first, so you pay for as few lookups as possible and send only to addresses that were confirmed.
Clay's Work Email waterfall guide describes it as a sequence that "cascades across multiple providers in sequence — stopping as soon as one returns a valid result." It starts with Infer Email, which guesses the address from the person's name and company domain and is "completely free." If that guess passes validation, "no paid providers are called."

Settings to check
Validation provider. Set one, so a validation step runs after each provider.
Require validation success. Turn it on. Clay says that when it is on, "the waterfall only accepts an email if the validation provider explicitly confirms it as valid."
Risk level. Clay offers Conservative, Balanced and Aggressive. Balanced includes catch-all addresses, which accept mail for any name and so can't be confirmed. For cold outbound, start with Conservative.
Provider order. Put the providers that do best on your market first. Test the order on a sample before running the whole table.
How do you keep Clay from re-enriching the same records?
Filter before you enrich, set each column to run only when the field it fills is empty or out of date, and turn off automatic re-runs on tables you aren't working on. Most wasted spend comes from running every column on every row, again.
Clay's own guide lists the same habits. It recommends you "Filter before enriching" so you don't waste "Actions and Data Credits on data you won't use," and that you "Use conditional logic" so enrichments "run only when a field is empty, data is outdated, or a confidence score is low." It also warns that "Re-running columns triggers new usage" and that Auto-run, which reruns columns when the table changes, should be off "when not needed."

A checklist for every table
One table per campaign, with the fit criteria written in its description.
Company enrichment and a filter column before any person lookup.
A run condition on every paid column: run only if the target field is empty.
Duplicates removed before import, matched on email and on company domain.
Auto-run off once the table is built.
A test run on a small sample, with the usage it consumed noted, before the full run.
If the same contacts also live in your CRM, clean them there first. Our guide to cleaning up CRM data covers duplicates and record ownership, so Clay doesn't enrich the same person twice under two spellings.
How do you send Clay leads without hurting deliverability?
Send in small batches from authenticated domains, only to validated addresses, and stop or slow down as soon as bounces or complaints rise. Clay prepares the list; the mailbox providers judge you on what happens after you send it.
Deliverability is whether your email reaches the inbox instead of spam. Gmail, Yahoo and Microsoft all set rules for bulk senders on authentication, one-click unsubscribe and complaint rates; we summarised them in our post on email sender requirements.
Before the first send
Every address passed validation under the Conservative setting.
The first batch is a few hundred contacts, reviewed by a person, not the whole table.
Each message uses a fact from enrichment that the reader would recognise as true.
Opt-outs and bounces flow back to Clay and the CRM so they are never re-imported.
Volume tools are tempting at this stage. Our post on AI SDRs covers where automated outreach helps and where it burns domains.
How we approach it at Vizio AI
We start with the fit, not the tool. Before building a Clay table we write the account and role criteria with the sales team and agree what a good lead looks like, then build company enrichment and filters before any person lookup.
We test every paid column on a small sample, put run conditions on each one, and record what a run consumes so the monthly budget is based on real usage. Results go to the CRM with a source field, so later reporting can tell which list and which workflow produced each meeting. This is part of how our Growth Studio connects outbound, CRM and reporting into one system.
Sources
Clay Docs, Actions & Data Credits: https://university.clay.com/docs/actions-data-credits
Clay, Compare plans, features & costs: https://www.clay.com/pricing
Clay FAQ, Do unsuccessful searches consume Data Credits? What about Actions?: https://www.clay.com/faq/do-unsuccessful-searches-consume-data-credits-what-about-actions
Clay Docs, Work Email waterfall: https://university.clay.com/docs/work-email-waterfall
Vizio AI, AI SDRs in 2026: What They Do Well and Where They Burn Your Domain: https://www.vizio.ai/resources/blog/ai-sdr
Vizio AI, Cold Email in Q4 2026: Authentication, One-Click Unsubscribe and Complaint Rates: https://www.vizio.ai/resources/blog/email-sender-requirements
Vizio AI, Clean Your CRM Data Before You Add AI: Dedupe, Grain and Ownership: https://www.vizio.ai/resources/blog/crm-data-cleanup
Vizio AI, Growth Studio: https://www.vizio.ai/studios/growth-studio
FAQs.
Read More

AI SDRs in 2026: What They Do Well and Where They Burn Your Domain

Cold Email in Q4 2026: Authentication, One-Click Unsubscribe and Complaint Rates

Clean Your CRM Data Before You Add AI: Dedupe, Grain and Ownership

Black Friday in Two Weeks: The Email and SMS Flows Worth Fixing Now

n8n vs Zapier vs Make in 2026: Picking an Automation Layer You Won't Rebuild

HubSpot vs GoHighLevel for Service Businesses Adding AI in 2026

The Power of B2B Marketing Automation for Business Success

Reading Invoices, Contracts and Forms With AI: Document Processing Without Retyping


