Clay · the GTM enrichment & orchestration engine
Clay is a programmable data layer between every source of truth in your stack and every place you act on data — 75+ providers, AI research agents, live buying signals, conditional activation. Treated as a point tool it burns credits; architected as infrastructure it feeds the entire funnel.
One table, four layers, a gate at every step — 10,000 raw records in
Every number on this page is computed, not asserted. The arithmetic is shown where it matters.
01 · What Clay actually is
Each column is an enrichment, an AI agent, an API call, a conditional formula, or a push into another system. Click a layer to see what it does and what it costs you when it's missing.
Changes: the layer detail shown below.
Layer 01 · Ingest
Pulls records from your CRM, CSVs, LinkedIn / Sales Nav, prospecting sources, webhooks, or Clay's own company and people search.
Skip it and: every audience lives in a different spreadsheet and nothing reconciles.
The mental model
Inputs → processes → outputs, with a constraint at every step. Clay is a controllable pipeline: every column is a step you can gate, so the system stays cheap and clean by design rather than by cleanup.
02 · The core engine · waterfall enrichment
Chain providers in order; a record matched by #1 never touches #2, so you never pay #2's credit on it. Move the inputs — the arithmetic below recomputes on your numbers.
Your inputs
Changes: whether all rows get enriched, or only the 62% that clear the ICP gate. Watch credit-per-verified-contact.
Credit price derived from Clay's published Growth plan: $495 / 6,000 Data Credits = $0.0825 per credit. Provider find rates are illustrative of Clay's 2025 work-email benchmark, where no single provider cleared both 95% quality and 90% coverage.
| Fall-through order | Records in | Found | Cumulative coverage | Credits |
|---|
Waterfallable
Work + personal emails, mobile and direct-dial phones, firmographics, technographics, LinkedIn profile and company data, titles and seniority.
Named providers
Prospeo, Dropcontact, Datagma, Hunter, People Data Labs, Nimbler, Apollo, Lusha, Snov — ordered best-accuracy-first, not cheapest-first.
The rule
Verify emails before they're allowed to leave the table. Unverified rows never reach a sequencer, so junk never touches a prospect.
03 · Claygent · AI research agents at scale
Point an agent at a company or person, give it a prompt, and it browses the live web — reading unstructured pages and returning a structured answer in a column. Set the two inputs below and the labor-versus-system comparison recomputes.
Your research load
What it returns
Leverage lens
Claygent converts a labor cost into a system cost — that is the whole thesis of removing the operator as the bottleneck, encoded in a column. 30% of Clay customers run it daily, ~500K research tasks per day across the base.
04 · Signals · the trigger layer
One signal is noise; stacked signals are intent. Toggle what this account is doing — the tier and the routing decision recompute live against the 3-signals-in-7-days threshold.
Signals detected in the last 7 days — changes: the account's tier and where it routes
Changes: whether signals are allowed to promote the account at all. Fit without timing is a static list; timing without fit is noise.
| Signal | What it tells you | Play it triggers |
|---|---|---|
| Funding round | New budget + growth mandate | Reach out while they're spending |
| Leadership hire | New VP/CxO reshaping their stack | New-exec change-agent outreach |
| Job change | A champion moved companies | Re-engage a warm relationship |
| Hiring / job posts | Team expansion, tooling gaps | Pitch the pain they're hiring for |
| Tech install / churn | They adopted or dropped a tool | Displacement or complement play |
| Web intent | They're researching your category | Prioritize accounts already in-market |
| Content engagement | Stakeholders on BOFU content | Route to sales as warm |
Two signals and a clean fit — message-matched sequence plus ad audience, not a rep's calendar.
Northwind Logistics · tier: ACCEPTABLE
Nurture track. Message-matched sequence queued; account synced to the ABM ad audience. No rep assignment yet.
signals: funding round, leadership hire
Configure the alert for 3+ signals in 7 days and route straight to a rep. That threshold is the difference between "big list, low reply" and "small list, booked meetings."
05 · Powering the full GTM motion
Outbound
Inbound & lead routing
CRM hygiene & enrichment
Marketing & ads
06 · Clay for ABM
Fit answers who the account is; signals answer what's happening. You spend attention only where the two overlap — which is the whole reason Clay's own ABM motion is operable rather than a slide.
Step 01
Define the ICP precisely
Firmographic + technographic + situational criteria, encoded as scored columns — not a vibe.
Step 02
Build the target list
Assemble and enrich candidates, gate to true ICP, dedupe against customers and open opps.
Step 03
Layer signals
Attach funding, hiring, job-change, tech and intent signals to every account.
Step 04
Score and tier
Combine fit + signal into weak / acceptable / strong priority tiers.
Step 05
Activate multi-channel
Strong accounts to a rep in Slack, synced ad audiences, message-matched sequences.
Step 06
Close the loop
Two-way CRM sync keeps engagement, ownership and opportunity state accurate — no account worked twice.
Tiering matrix · fit × signals in 7 days
Bands are the diagnosis, not decoration: an account only reaches a rep's calendar in the top-right cell.
Benchmark for "strong"
Tier a target account as high-priority when it clears the ICP fit gate and shows 3+ situational or intent signals inside a 7-day window.
Alignment beats volume. The list gets smaller and the calendar gets fuller.
07 · The integration stack
Native, mostly two-way integrations across the stack, plus an HTTP API for anything custom.
| Category | Integrations | Role in the system |
|---|---|---|
| CRM | HubSpot, Salesforce | Source of truth; two-way enrich + sync |
| Sequencers | Smartlead, Instantly, Outreach, Salesloft, Apollo, HubSpot Sequences | Execution layer for verified contacts |
| Alerting | Slack | Real-time signal + routing notifications to reps |
| Data sources | 75+ providers, LinkedIn / Sales Nav, web scraping | Coverage for the waterfalls |
| Automation | Zapier, webhooks, HTTP API | Glue for custom flows and internal tools |
| Ads | Ad-audience syncs | ABM targeting from enriched segments |
The canonical flow
The critical design rule: make destinations conditional. Only rows that passed the ICP gate and email verification fire the HubSpot or Smartlead action, so junk never leaves the table and never touches a prospect.
08 · The Clay MCP · agentic GTM
The MCP server exposes Clay to AI assistants as callable tools, so prospecting, enrichment and table operations happen in plain language instead of clicks.
$ claude mcp add clay \ --transport http https://api.clay.com/v3/mcp > OAuth authorized · credentials reused > tools: search_companies, search_people, enrich_record, read_table, write_table you find 50 fintech companies in NYC that raised a Series B this quarter, enrich decision-makers, load them into my outbound table agent 50 accounts matched → 187 contacts enriched → 164 verified → written to "Q3 Outbound"
| Type | Official hosted remote MCP server |
| Endpoint | https://api.clay.com/v3/mcp |
| Transport | Streamable HTTP (current MCP standard) |
| Auth | OAuth — authorize once in-browser, credentials reused |
| Clients | Claude / Claude Code, Cursor, VS Code, Codex, other MCP-compatible clients |
Why it changes the operator's job
The MCP collapses the interface. Instead of a person building tables click-by-click, an agent orchestrates Clay directly — and chains it with your CRM, Slack and calendar in one instruction. Same move as replacing manual delivery with a system, applied to the data layer itself.
09 · Use-case catalog
Changes: which plays are shown below.
Foundational
CRM enrichment & cleanup
Fill gaps, standardize fields, dedupe, re-enrich on a schedule.
Foundational
Verified-contact list building
Waterfall emails and phones for any target segment.
Foundational
Inbound enrichment & instant routing
Turn a form-fill into a scored, routed lead in seconds.
Signal-driven
Job-change plays
Chase champions when they move to new companies.
Signal-driven
Funding plays
Reach newly funded accounts while the budget is fresh.
Signal-driven
New-hire plays
Target change-agent execs inside their first 90 days.
Signal-driven
Tech-install displacement
Target companies that just adopted a competing or complementary tool.
AI · Claygent
Auto-qualification
Score every account against ICP with a written reason attached.
AI · Claygent
Hyper-personalized first lines
One true, specific detail per prospect — sourced, not invented.
AI · Claygent
Competitive & market research
Scrape reviews, spot trends, enrich supplier lists with live attributes.
ABM & advanced
Signal-stacked account lists
3+ signals in 7 days routed to a named rep in Slack.
ABM & advanced
Enriched ad audiences
Deduped, ICP-gated segments synced for account-based targeting.
ABM & advanced
Agentic GTM via MCP
An assistant builds, enriches and activates lists on command.
10 · The JayOh implementation playbook
The discipline below is what separates a runaway credit bill from a compounding revenue engine. Build the gates before you spend a single credit.
Phase 1 · Architect
Before any enrichment
Phase 2 · Gate every step
Spend nothing on records you don't want
Phase 3 · Enrich in waterfalls
Order matters more than vendor count
Phase 4 · Activate conditionally
Nothing fires unconditionally
Phase 5 · Operate & compound
The system runs without you
Schedule re-enrichment and signal monitoring so the data stays fresh with no operator in the loop.
Review find rates, credit-per-qualified-record and reply rates monthly; retune the waterfall order.
Layer the MCP so an agent can extend and run the system on command.
11 · Credit economics & cost control
After the March 2026 overhaul, Clay bills two credit types: Data Credits for buying enrichment data, Actions for running workflows and routing.
Operator dashboard · monthly
Credit efficiency, not credit spend
Find rate by field
Credit / qualified record · 6 months
Gates added in month 3, waterfall reordered in month 5. Same providers, same list — 65% lower cost per qualified record.
| Clay plan (2026) | Price | What you get |
|---|---|---|
| Free | $0/mo | Trial the platform, limited credits |
| Launch | $185/mo | ~2,500 Data Credits, 15,000 Actions, phone enrichment, signal tracking |
| Growth | $495/mo | ~6,000 Data Credits, 40,000 Actions, CRM sync, HTTP API, Web Intent, Ads, priority support |
| Enterprise | Custom | Higher volume, security and support |
Cost-control checklist
CRM sync and the HTTP API land on Growth and up, which is the practical floor for a real RevOps deployment. Figures move — confirm current numbers on clay.com/pricing before quoting.
12 · Common failure modes
| Failure mode | The fix | Severity |
|---|---|---|
| Enrich-first, think-later — credits burned on junk | Build ICP + dedupe + disqualification gates before any enrichment. | Critical |
| Single-provider mindset — low coverage | Use waterfalls; order best-accuracy-first. | Critical |
| Unconditional pushes — bad data hits prospects and CRM | Make every destination conditional on gates + verification. | Critical |
| Static lists — data decays, timing missed | Schedule re-enrichment and signal monitoring. | High |
| Treating Clay as a tool, not infrastructure | Architect it as the data layer feeding the whole funnel. | High |
| No efficiency metric | Track credit-per-qualified-record and reply rate; tune monthly. | High |
13 · Where JayOh fits
A Clay implementation is a RevOps build: ICP encoded as logic, waterfalls tuned for coverage and cost, signals wired to routing, CRM sync kept clean — handed off as a system that runs without the operator in the loop.
Clay turns messy GTM reality into a structured, self-updating revenue engine. Architected well, it removes the operator as the bottleneck without lowering the standard: more coverage, better timing, cleaner data, less manual work.
Next moves
JayOh · We Engineer Growth Systems · Joe@Jayoh.io
We engineer growth systems. JayOh builds the RevOps architecture, GTM execution and data infrastructure that turn messy revenue operations into predictable pipeline for B2B SaaS and PE-backed teams.
