How to Automate Inbound Lead Qualification in Clay: A Step-by-Step Guide

Reaching a lead in the first few minutes only helps if you have something to say when they pick up, and a name plus an email is not something to say. The teams that win inbound do both at once: they respond in minutes and they respond with context, on every lead, automatically. Speed is easy to mandate. Context is harder. This is how to build both into one workflow.

What you need before you start

A Clay account and your form tool (Typeform, Webflow, HubSpot forms, or anything that supports webhooks). A working ICP hypothesis, even an imperfect one. A sequencer (Instantly, Smartlead, or Clay's native Sequencer) and a CRM for the export step. You refine scoring criteria as the workflow runs. You do not need a perfect model before you begin.

Step 1: Connect your inbound lead sources to Clay

Every inbound source pipes into Clay the same two ways: native integration or webhook. Both deliver leads to a Clay table the moment a form is submitted.

For Typeform, create a new table in Clay, go to Import, choose Import Typeform form responses, and link your account. New submissions appear automatically as they arrive.

For every other form tool, go to Import in Clay, select Pull in data from a Webhook, copy the webhook URL, and paste it as a new webhook destination in your form tool. Webflow forms, HubSpot forms, Intercom, Zoom webinar registrants, trial signups, chatbot completions: any tool that can send an HTTP POST feeds a Clay table in real time.

Clay tables have a 50,000-row limit across all plans, which applies to every source, so rotate to a fresh table before a high-volume webhook fills one. If you run high-volume lead programs, build a table-rotation system: a new table starts automatically when the current one reaches capacity, or segment by lead source so each table handles one channel's volume.

One nuance on lead source overlap: a trial signup and a demo request from the same email address in the same week are not two separate leads. They are two intent signals from one buyer. Cross-reference by email domain before enrichment and merge duplicates with a Clay formula so both signals appear on one record.

Step 2: Enrich every submission from a single email address

Your inbound form only needs to ask for name and email. Nothing else. This is the most counterintuitive step for teams new to Clay.

Before Clay, teams faced a genuine tradeoff. Longer forms meant better data for qualification. Shorter forms meant higher completion rates but reps scrambling to research leads by hand. Every extra field you add to a form costs you completions. Teams picked between conversion rate and data quality, and usually lost on both.

From name and email, Clay enriches in two parallel tracks. One track works the company: it matches the email domain to a company record, then pulls headcount, industry, tech stack, funding stage, and description. The other track works the person: job title, seniority, time in role, and professional background. Both run at the same time.

For email addresses from free providers like Gmail, Yahoo, or personal Outlook, do not disqualify automatically. A Head of Engineering at a Series B company sometimes submits from personal email, especially on mobile. Run name alongside email through enrichment and let the scoring model decide, not the email domain.

Use a multi-provider waterfall for email verification on any contact addresses you find. Findymail first for quality, Wiza for coverage, Enrow for cost. Clay charges credits only when a provider returns a verified result. A waterfall that stacks several providers covers far more of a list than any single provider can on its own, and you only pay for the hits.

Step 3: Understand what enrichment actually does to a lead record

Enrichment is not "filling in some fields." A name and email go in, and a record a rep can act on without any research comes out. Most teams underestimate how far apart those two things are.

Inbound enrichment runs the opposite direction from outbound. With outbound, you start from a company and build out before finding contacts. With inbound, you start from a specific person who just raised their hand. Begin with the email, resolve company and contact at the same time, then go deeper on whatever signals your scoring model needs.

Step 4: Score leads with criteria that match your ICP

A score of 84 does not mean "almost perfect." It means this lead has the specific traits your team has learned to associate with conversion. A scored list tells reps exactly where to start, and the weights behind that score are yours to define.

Build the score in a Formula column. Add a column, set the type to Formula, and use the Formula Generator to total points across attributes like company size, seniority, and tech-stack fit, or to assign letter grades from combined criteria. For signals that need judgment rather than a lookup, add an AI formula column:

Test on 20 rows before running at scale. Read every output. If a meaningful share look generic or contradictory, tighten the prompt before you proceed.

Step 5: Route each lead to the right rep, automatically

Scoring sets priority. Routing decides who handles it. The two do not have to follow the same rule.

Clay's Weighted Round Robin handles most routing cases. Build a rep table with a weight column and Clay assigns leads in proportion to each rep's weight; use dynamic lists so the rotation adjusts as reps are added, removed, or marked out of office. If rep data lives in Salesforce, import the table from SFDC on a scheduled refresh so it stays current as team structure changes.

For Hot leads, routing needs to do more than assign. It needs to alert. The rep should not have to discover that a Hot lead appeared. The notification should come to them.

For territory-based routing, add a territory column to the rep table and a lookup formula that matches the lead's company region to the correct rep before round-robin runs. Territory lookup fires first; round-robin applies within the matched territory only.

Step 6: Generate personalized follow-up before the rep logs in

A rep who calls four minutes after form submission and opens with "Hi, I saw you filled out our form" has speed and nothing else. The first message is where the homework shows.

Before writing any follow-up, run a CRM lookup. Use the Lookup Record enrichment to check whether this contact or company already exists. If there is an active deal, route to a different workflow. Reaching out to someone already mid-negotiation with another rep is the fastest way to lose credibility and rep trust in the system.

Then generate one AI-written opening line per lead from the enrichment data already in the table. One sentence, not a paragraph.

Run this prompt on every Hot and Warm lead in the table. The output plugs directly into your sequencer as the email's first line. The rest of the sequence is templated. Only the opener is personalized. That one sentence, sent within minutes, is what makes the follow-up feel human.

Step 7: Export to CRM and trigger outbound sequences

An enriched, scored, routed record is worth nothing until it lands in the tools your team actually opens. The export step writes everything to its final destination.

For the CRM, run a lookup by email before creating any record. If a contact exists, update it with the new enrichment data. If not, create a new contact and company record with all enriched fields mapped to the right CRM fields. Build custom fields that match your scoring dimensions (ICP Tier, Key Signal, Seniority Tier) before the sync runs. Standard CRM fields do not surface the right context for reps.

For sequence enrollment, use a conditional action immediately after the CRM write:

• Hot leads (75+): Enroll immediately; the personalized opener fires as the first step.
• Warm leads (50 to 74): Enroll in a 5-day SDR cadence with a 24-hour delay.
• Nurture leads: Create the record only and hold 30 days.
• Disqualified: Write the AI qualifier's reasoning as a CRM note and stop.

For the Slack digest, go beyond individual Hot lead alerts and build a morning summary that fires at 8am for each rep, covering their previous 24 hours of leads: name, score, company, one-line qualifier output, link to CRM queue. Reps start every day knowing which three calls to make first.

What teams running this workflow achieve

Common failure modes

• Skipping the CRM duplicate check before routing: Enrolling a lead who is already in an active deal is the fastest way to break rep trust in the system. Run the lookup before every routing and sequence action.
• Scoring before normalizing: Seniority rules that look for VP miss Vice President on raw enriched data. Normalize company names and job titles before scoring runs on every row.
• Single enrichment provider for email: Single-provider coverage reaches 60 to 75% of leads. The 25 to 40% that fail to enrich never get scored or routed. A waterfall adding two more providers takes 20 minutes to configure and frequently doubles coverage.
• Running AI formulas without testing 20 rows first: Personalization prompts that seem tight in isolation often produce generic output at scale when edge-case inputs arrive. Read the first 20 outputs manually before enabling at volume.
• Routing to reps with no Slack alert: CRM record creation alone is not enough for Hot leads. Reps check their queue on their own schedule. The Slack alert is what creates the sub-5-minute response.