how to use clay for enrichment

Clay enrichment refers to the process of using Clay (clay.com) to automatically append, verify, and synthesize data on leads, accounts, or contacts — pulling from 75+ integrated data providers in a single workflow. Unlike point-solution enrichment tools (Clearbit, ZoomInfo, Apollo alone), Clay acts as an orchestration layer: you define what data you need, which sources to try, and in what order — and Clay handles the rest.

The reason Clay outperforms single-source enrichment is **coverage + freshness**. No single provider owns 100% of B2B contact data. ZoomInfo is strong on enterprise, Apollo on SMB, Datagma on EU — but gaps are universal. Clay's waterfall model means if Source A doesn't return a mobile phone, Source B gets triggered automatically. You pay only for successful matches on most integrations, so cost efficiency scales with result quality.

Clay also handles **unstructured enrichment** through Claygent, its AI web agent. Instead of just querying databases, Claygent can scrape a company's careers page, summarize recent news, identify tech stack signals from job descriptions, or extract pricing tiers from a website — enrichment that no static database can provide.

For GTM teams, this matters because personalization at scale requires more than name, title, and email. You need buying signals, pain point indicators, company growth data, and tech stack context to write outreach that converts. Clay is the infrastructure that makes that possible without a data engineering team.

The 4 S's of Clay is a practitioner framework for structuring enrichment workflows effectively. Whether you learned it from a Clay power user, a RevOps Slack community, or derived it from working with the tool, the framework maps cleanly to how Clay actually operates:

**1. Sources** — Where your data comes from. Clay integrates with 75+ providers including Apollo, Clearbit, LinkedIn (via RocketReach, Proxycurl, or PhantomBuster), Hunter.io, Datagma, Lusha, and more. Choosing the right sources for your ICP is step one. If you're targeting European companies, lean on Datagma and Kaspr. For US enterprise, ZoomInfo or Clearbit. For SMB and startup, Apollo and LeadMagic.

**2. Signals** — What you're actually trying to learn about a contact or company. Signals go beyond basic firmographics. Job postings signal hiring pain and budget. Funding rounds signal growth and tech investment appetite. Tech stack data signals tool overlap or displacement opportunities. Clay lets you build signal-rich records by combining Claygent web scrapes with database lookups.

**3. Sequencing** — The waterfall logic that defines which source runs first, second, and third for each data point. Sequencing is where most teams underinvest. A well-sequenced waterfall minimizes credit spend, maximizes match rates, and ensures fallback coverage. Example: for email, try Apollo → Hunter → Datagma → Claygent scrape → mark as unknown.

**4. Synthesis** — Using Clay's built-in AI (GPT-4o integration) to turn enriched data into outputs: personalized email icebreakers, account summaries, lead scores, or segment tags. Synthesis is what converts raw enrichment into revenue-ready assets. A common output is a one-line custom first sentence per prospect built from their LinkedIn bio, company news, and job title — generated at scale.

Here's how to build your first Clay enrichment workflow from scratch:

**Step 1: Create a Clay Table** Start a new table in Clay. This is your working spreadsheet. You can populate it by importing a CSV, connecting a CRM (HubSpot, Salesforce), syncing from Apollo, or using Clay's built-in prospecting to generate a list from filters (company size, industry, geo, tech stack).

**Step 2: Add Enrichment Columns** Click the '+' to add a new column and search for the enrichment action you want — 'Find email,' 'Get LinkedIn profile,' 'Enrich company,' etc. Clay will prompt you to connect the relevant provider and map your input field (usually name + company domain).

**Step 3: Build Your Waterfall** For critical fields like email or phone, don't rely on one provider. Stack multiple enrichment columns for the same data point, then use a 'Fallback' or 'Conditional' column to return the first non-null result. This is the waterfall. A typical email waterfall: Apollo → Hunter.io → Datagma → Claygent → null.

**Step 4: Add Claygent for Signal Enrichment** For intent and context signals, add a Claygent column. Give it a URL (company website, LinkedIn, G2 page) and a specific prompt: 'What does this company's job postings tell us about their data infrastructure challenges?' Claygent returns structured text you can feed into your synthesis step.

**Step 5: Run AI Synthesis** Add a GPT-4o column. Reference enriched fields as variables and write a prompt: 'Using {{job_title}}, {{company_description}}, and {{recent_news}}, write a one-sentence personalization for a cold email.' Output populates instantly across all rows.

**Step 6: Push to Your Sequencer** Use Clay's native integrations to push enriched, personalized records directly to Instantly, Smartlead, Outreach, or HubSpot. Map fields to your sequence variables and launch.

Clay connects to 75+ data providers, but not all are equal for every use case. Here's a practitioner breakdown of which sources to prioritize:

**For Email Addresses:** - **Apollo.io** — Best coverage for US SMB and mid-market. Good for bulk email finding. - **Hunter.io** — Strong for finding pattern-matched emails when domain is known. - **Datagma** — Strong EU coverage, good for GDPR-compliant email sourcing. - **LeadMagic** — Excellent for LinkedIn-sourced email finding. - **ZeroBounce / NeverBounce** — Use these *after* finding emails to verify deliverability before sending.

**For Phone / Mobile:** - **Lusha** — High mobile hit rate in North America. - **Kaspr** — Strong for European mobile numbers. - **Proxycurl** — LinkedIn-sourced, good for direct dials from profiles.

**For Company Firmographics:** - **Clearbit** — Best-in-class for company data enrichment (revenue estimates, headcount, tech stack). - **Apollo** — Adequate firmographics, useful when already using Apollo for contacts. - **Harmonic** — Best for startup/VC-backed company funding and growth signals.

**For Intent and Signal Data:** - **Bombora** — Topic-level intent data for enterprise ABM. - **BuiltWith / Wappalyzer** — Tech stack detection via Clay's HTTP request action. - **Claygent (native)** — Best for unstructured, bespoke signal scraping.

**For LinkedIn Data:** - **Proxycurl** — Most reliable LinkedIn profile enrichment. - **RocketReach** — Good fallback for LinkedIn-sourced emails and profiles. - **PhantomBuster** — Useful for LinkedIn list scraping, though requires more setup.

Match your source selection to your ICP geography, company stage, and data freshness requirements. Always waterfall — never rely on a single source for mission-critical fields.

In the context of Clay (the GTM tool), 'Vaseline' is a colloquial term used in some practitioner communities to describe **lubricant logic** — small configuration tweaks that reduce friction, prevent workflow breakage, and make enrichment run more smoothly at scale. It's an analogy borrowed from physical clay work, where Vaseline (petroleum jelly) is applied to molds and tools to prevent sticking and enable clean separation.

In Clay workflow terms, 'Vaseline' techniques include:

**1. Conditional Logic as Friction Reducer** Using 'Run if' conditions so enrichment columns only trigger when upstream data is present. This prevents wasted credits on rows where required inputs (like a LinkedIn URL or company domain) are missing. Example: 'Only run Claygent if company_website is not empty.'

**2. Deduplication Before Enrichment** Running a dedup step on your input list before triggering any enrichment. Duplicate rows waste credits and inflate your enriched list size. Use Clay's built-in dedup or a pre-processing formula column.

**3. Null-Handling Formulas** Building fallback formulas so that downstream synthesis columns don't break when enrichment returns null. A simple COALESCE-style formula: 'If {{email}} is empty, skip this row in the export.' This prevents garbage data from entering your sequencer.

**4. Rate Limit Management** Some providers (Proxycurl, Claygent-heavy workflows) can hit rate limits mid-run. Setting row-level processing delays or running in batches is the 'Vaseline' that keeps large tables from stalling.

Think of Vaseline logic as defensive engineering — it doesn't add data, but it ensures the data you do add flows cleanly through the workflow without causing downstream problems.

Clay enrichment looks different depending on where in the GTM motion you're applying it:

**Outbound SDR Workflows** The primary use case. Build a list from Clay's prospecting filters or import from Apollo/LinkedIn Sales Navigator. Run a waterfall for email + phone. Add a Claygent column scraping each company's website for a personalized pain point signal. Use GPT-4o to generate custom first lines. Push to Instantly or Smartlead. This workflow reduces research time per prospect from 10+ minutes to under 10 seconds at scale.

**Account-Based Marketing (ABM)** For ABM, enrichment depth matters more than volume. Use Clay to build a single 360° view per target account: firmographics (Clearbit), funding (Harmonic), tech stack (BuiltWith), open roles (Claygent on careers page), recent news (Claygent on Google News), and all buying committee contacts with their LinkedIn summaries. Export a structured account brief that AEs use for personalized outreach and discovery calls.

**RevOps / CRM Hygiene** Many RevOps teams use Clay to continuously enrich and update their CRM. Connect Salesforce or HubSpot, pull records with stale data, run enrichment to refresh titles, emails, phone numbers, and company data, then push updated values back. This solves the 'dirty CRM' problem without manual research. Schedule recurring Clay table runs monthly to maintain data freshness.

**Inbound Lead Scoring** When a new lead fills out a form, trigger a Clay webhook enrichment: instantly pull company data, LinkedIn profile, tech stack, and funding signals, then score the lead based on ICP fit. Push the scored lead to your CRM with a priority flag. This compresses inbound response time and gives AEs context before the first call.

Even experienced operators make these mistakes when building Clay enrichment workflows:

**1. Running All Providers Simultaneously (No Waterfall)** Triggering Apollo, Hunter, Datagma, and Lusha all at once on every row burns credits fast. Build a sequential waterfall and stop at the first successful match.

**2. Not Validating Emails After Finding Them** Found emails ≠ deliverable emails. Always run a verification step with ZeroBounce or Neverbounce after email enrichment, especially before sending to Instantly or Smartlead. Bounce rates above 5% kill domain reputation.

**3. Using Claygent Without Specific Prompts** Vague Claygent prompts return vague output. 'Tell me about this company' is useless. 'List the top 3 technical challenges mentioned in their last 3 job postings' returns actionable signal.

**4. Ignoring Credit Costs on Large Tables** Clay credits add up fast on tables with 5,000+ rows and 10+ enrichment columns. Audit your waterfall logic and conditional triggers before running large batches. Use Clay's credit estimator.

**5. Over-Enriching for the Use Case** You don't need 30 data points per lead for a cold email. Identify the 3-5 fields that actually drive personalization and conversion, and enrich only those. More data ≠ better outreach.

**6. Skipping Deduplication** Importing lists from multiple sources (Apollo + LinkedIn + CRM export) without deduping inflates your table and wastes credits on duplicate contacts. Always dedup on email or LinkedIn URL before enriching.

Enrichment data decays. Job titles change, companies get acquired, emails go invalid — B2B data has an estimated 30% annual decay rate. Here's how to maintain the value of your Clay-enriched data:

**Timestamp Every Enrichment Run** Add a 'Last Enriched' date column to every Clay table. This lets you filter and re-enrich records older than 90 days on a rolling basis rather than re-enriching your entire database.

**Use CRM as Your System of Record** Don't treat Clay tables as permanent storage. Push enriched data to Salesforce or HubSpot with clear field mapping. Clay is a processing layer, not a database. Treat it accordingly.

**Schedule Recurring Enrichment Runs** For active prospects and customers, set a monthly or quarterly Clay workflow to re-enrich key fields (title, email, phone, company headcount) and update your CRM automatically.

**Archive, Don't Delete** Instead of deleting stale or bounced leads from your Clay tables, add a 'Status' column and mark them as 'invalid' or 'archived.' This preserves your audit trail and prevents accidentally re-enriching the same bad data.

**Monitor Bounce Rates as a Decay Signal** If your Instantly or Smartlead campaign bounce rates start climbing above 3-4%, that's a signal your enrichment data is aging. Trigger a re-verification pass through ZeroBounce on that segment.