A complete setup guide for account scoring, enrichment waterfalls, and automated research pipelines.
TL;DR
Clay is an enrichment and automation platform -- not just a data tool.
Define your ICP before building your first table. Clay amplifies your targeting -- good or bad.
Build an enrichment waterfall: multiple providers, sequential fallback, one verified email.
Use Claygent for AI-powered account research that goes beyond structured data.
Connect Clay to your sequencer via webhook. The pipeline should be fully automated.
Clay is a prospecting and enrichment platform that connects dozens of data providers into a single workflow. It doesn't just find contacts -- it researches them, scores them, and routes them to the right outreach sequence. Used well, it's the core of an agentic outbound system. Used poorly, it generates expensive garbage faster.
What Does Clay Actually Do?
Most people use Clay as a contact finder. That's a fraction of what it does.
Enrichment: pull data from 75+ providers in a single workflow
Waterfall logic: try Provider A, fall back to B, then C -- stop when you have what you need
AI research: Claygent browses the web and synthesizes account intelligence
Scoring: build formulas that score accounts on fit and signal
Automation: trigger workflows based on CRM events, webhooks, or schedules
Sequencer integration: push enriched, scored contacts directly to your email tool
Think of Clay as the operating system for your outbound stack. Everything flows through it.
What Should You Define Before Touching Clay?
Clay amplifies your targeting decisions -- good or bad. If your ICP is wrong, Clay will find the wrong people faster and cheaper.
Company size: headcount range and revenue range
Industry: be specific -- not 'tech' but 'B2B SaaS with sales teams over 10 reps'
Geography: which countries or regions
Tech stack signals: what tools do your best customers use
Exclusions: what company types should never enter your pipeline
Contact titles: the 3–5 titles most likely to own your problem
Document this before you open Clay. It becomes your scoring model.
How Do You Build a Clay Table from Scratch?
1Create a new table -- this is your prospecting workspace
2Import your seed list: companies from your CRM, a CSV, or a LinkedIn Sales Nav export
3Add a domain column and normalize it -- everything downstream depends on a clean domain
4Run company enrichment first: headcount, revenue, industry, location
5Filter rows that don't match your ICP before spending credits on contact finding
6Run contact enrichment for qualifying companies: find decision-makers by title
7Run email enrichment: waterfall across providers to maximize coverage
8Run email verification: only deliverable emails proceed
Credit discipline rule
Enrich company data first -- it's cheap
Filter aggressively before running contact enrichment -- it's expensive
Only verify emails that passed all filters -- verification costs stack up
Never run Claygent on accounts you haven't pre-qualified
How Do You Build an Enrichment Waterfall?
An enrichment waterfall queries multiple data providers in sequence. If Provider A finds the email, stop. If not, try Provider B. This maximizes coverage while minimizing cost.
Step
Provider type
Fallback trigger
1
Primary email finder (e.g., Apollo, Hunter)
No result or low confidence
2
Secondary finder (e.g., Findymail, Icypeas)
Still no verified result
3
LinkedIn-based finder
Still no result
4
Email verification
Always -- run on every found email
5
Block if undeliverable
bounce risk > threshold
A well-built waterfall gets 70–85% email coverage on a clean ICP list. Under 60% usually means your ICP includes too many small or obscure companies.
70–85%
email coverage
Typical yield from a properly built enrichment waterfall on a well-defined ICP list.
How Do You Build an Account Scoring Model in Clay?
Account scoring in Clay uses formula columns. You assign points to signals and sum them into a score. Accounts above the threshold advance to outreach.
Headcount in ICP range: +20 points
Revenue in ICP range: +15 points
Uses a relevant tool in their tech stack: +20 points
Recent funding event (last 6 months): +15 points
Relevant job posting active: +15 points
Industry exact match: +10 points
Geography match: +5 points
Threshold: 60+ points advances to outreach. 40–59 goes to a review queue. Under 40 is excluded.
Adjust weights based on what your best customers actually look like. The model should reflect your real win rate signals -- not gut instinct.
What Is Claygent and When Should You Use It?
Claygent is Clay's AI browsing agent. It can visit any URL, extract information, and answer freeform questions about a company.
Summarize a company's main product from their homepage
Find the most recent press release or news mention
Identify what problem they solve and who they serve
Extract pricing page information
Find the LinkedIn headline of a specific contact
Determine if a company is a competitor, partner, or prospect
Claygent costs more credits than structured enrichment. Use it only after accounts have passed scoring. Don't run it on your full table.
1 Claygent prompt
per qualified account
Run Claygent after scoring to add a personalization hook -- never before. Credit cost is too high to run on unfiltered lists.
How Do You Connect Clay to Your Sequencer?
Clay pushes contacts to your sequencer via webhook. When a row meets your export criteria -- verified email, score above threshold, no CRM match -- Clay fires the webhook automatically.
1Create a view with your export filter (score ≥ 60, email_status = deliverable, crm_match = false)
2Add a webhook column -- Clay fires this when the row enters the view
3Map fields: email, first name, company, and any custom variables your sequence uses
4Add a 'pushed' boolean column -- prevents duplicate enrollments if Clay re-evaluates the row
5Test with 3–5 rows before enabling for the full table
What Are the Most Common Clay Mistakes?
Running enrichment before filtering -- you'll spend 3x more credits than necessary
Skipping email verification -- unverified emails bounce and damage your domain
Not deduplicating against your CRM -- you'll email existing customers or open opportunities
Using Claygent on the full table -- reserve it for accounts that passed scoring
No 'pushed' gate on the webhook -- contacts get enrolled multiple times
Treating Clay as a one-time export -- it should run continuously as new accounts qualify
Building everything in one table -- separate your enrichment, scoring, and export logic
What Does a Production Clay Workflow Look Like?
Table
Purpose
Trigger
Account intake
Company enrichment + ICP filtering
Daily import from signal sources
Contact finder
Find decision-makers at qualifying accounts
Account passes ICP filter
Email waterfall
Verify deliverable email for each contact
Contact found
Scoring
Score accounts on fit + signal
Email verified
Claygent research
Pull personalization hook
Score ≥ 60
Export queue
Push to sequencer via webhook
All gates passed
This structure runs continuously. New signals enter the top. Qualified, researched, verified contacts exit the bottom into your sequencer. Your reps see only contacts that passed every gate.