Claude Enrichment: How to Enrich B2B Contacts Inside Claude (2026)
By Kushal Magar · July 6, 2026 · 11 min read
Key Takeaway
Claude enrichment turns Claude Code into your enrichment tool. Connect SyncGTM's MCP server and ask in plain English — Claude pulls 76+ verified data points per contact (work email, mobile, LinkedIn, tech stack, funding, and intent signals) and writes them straight to your file. No exports, no tab-switching, no per-seat enrichment app.
TL;DR
- Claude enrichment means enriching B2B contacts from inside Claude Code — no separate enrichment app, no manual CSV exports.
- You connect an enrichment MCP server, then ask Claude in plain English to enrich a contact, a company, or a whole list.
- With SyncGTM's MCP, one prompt can pull 76+ data points per contact: verified work email, mobile, LinkedIn, tech stack, funding, and intent signals.
- Setup is one command. This guide walks the full workflow, from install to batch-enriching a CSV.
- The data layer is what makes it work — a waterfall across 50+ providers is the difference between a 40% fill and an 85% fill.
Overview
Enrichment used to mean leaving your workflow. You exported a list, uploaded it to Clay or Apollo, waited, downloaded the result, and merged it back by hand.
Claude enrichment collapses that loop. You stay in Claude Code, describe what you need, and Claude calls live enrichment tools through MCP — returning verified data in the same conversation.
This tutorial is for founders, SDRs, and GTM engineers who want a repeatable enrichment workflow inside the terminal. You will install the tools, enrich a single contact, then scale it to a full CSV.
We use SyncGTM's MCP server as the data layer throughout, because it exposes person, company, tech-stack, and signal tools in one install. The workflow itself applies to any MCP enrichment provider.
What Is Claude Enrichment?
Claude enrichment is the practice of enriching contacts and companies from inside Claude Code, using MCP tools instead of a standalone enrichment platform. Claude becomes the interface; a data provider becomes the engine.
The mechanism is MCP (Model Context Protocol), the open standard that lets Claude call external tools. An enrichment MCP server registers tools like enrich_person or find_work_email, and Claude calls them like any other action.
The difference from a normal enrichment app is that there is no UI to click through. You type a request in English, Claude picks the right tools, runs them, and writes the results wherever you want them.
It is the enrichment step of a full Claude GTM motion, isolated. Once it works on one contact, the same pattern scales to a list and slots into your outreach.
Why Enrich Contacts Inside Claude?
Reps do not spend their week selling. According to Salesforce's State of Sales, sellers spend roughly 70% of their time on non-selling work — a large share of it research and data entry.
Traditional enrichment adds friction to that tax. Every export, upload, and re-merge is a context switch, and every context switch is a place the workflow stalls.
Claude enrichment removes the switching. The data lands where you already are — next to the file you are building, the email you are drafting, the CRM record you are about to update.
It also composes. Because Claude holds context across steps, enrichment feeds straight into personalization and cold email without you moving a single row between tools.
The 76+ Data Points You Can Pull Per Contact
With SyncGTM's MCP, Claude enrichment returns 76+ data points per contact, grouped into four layers. You rarely need all of them — you pull the ones your outreach actually uses.
| Layer | Example data points | SyncGTM MCP tools |
|---|---|---|
| Person | Verified work email, personal email, mobile, work phone, job title, seniority, LinkedIn URL | enrich_person, find_work_email, find_mobile_number, verify_email |
| Company | Industry, headcount, revenue band, funding, HQ location, domain | enrich_organization |
| Technographics | The exact tools and platforms a company runs | find_company_techstack |
| Intent signals | Job changes, promotions, hiring surges, product launches | check_job_change, check_promotions, company_job_openings, company_product_launch |
That range is the point. A single-source tool gives you an email and a title; Claude enrichment stacks person data, firmographics, tech stack, and timing signals onto the same record.
For a deeper look at the company side specifically, see our guide to pulling firmographic data inside Claude Code.
What You Need Before You Start
Three things, and you likely already have the first. Setup takes about five minutes end to end.
- Claude Code. Installed and signed in, via Claude Max or a Team seat. This is where every command runs.
- A SyncGTM account. The free plan is enough to test the full workflow — no credit card required.
- A contact to enrich. A name plus a company, an email, or a LinkedIn URL. Any one of these is enough to start the waterfall.
Step 1: Install the SyncGTM MCP Server
Add the SyncGTM MCP server to your Claude Code environment with one command:
claude mcp add syncgtm -- npx -y syncgtm-mcpThen type /mcp in Claude Code and confirm SyncGTM shows as active. If it does, all the enrichment tools are now available to Claude.
Run check_credits once to confirm your balance is connected. The free plan gives you enough credits to enrich a test batch before you commit.
Step 2: Enrich a Single Contact
Start with one person so you can see exactly what comes back. Give Claude whatever identifier you have.
Enrich this contact with SyncGTM:
Name: Jordan Lee
Company: Northbeam
Use enrich_person. Return job title, seniority,
LinkedIn URL, and company info.Claude calls enrich_person — which resolves a person by name, email, or LinkedIn URL — and returns contact details, job title, and company info in the conversation.
No form, no CSV upload. You read the result inline and decide what to pull next.
Step 3: Add Waterfall Email, Mobile, and Verification
A profile without a reachable email or number is not send-ready. This is where the waterfall does its work.
For the same contact:
1. find_work_email
2. find_mobile_number
3. verify_email on the result
Return only verified, deliverable contact details.SyncGTM runs each lookup across 50+ providers. If the first source has no match, it falls through to the next, which is what pushes fill rates toward 85% instead of the ~40% you get from a single source.
The final verify_email call checks deliverability before you ever hit send, so nothing bounces. If you want the theory behind this, read what waterfall enrichment is and why it beats single-source data.
Step 4: Enrich the Company Behind the Contact
A person is only half the record. Company context is what lets you write outreach that lands.
Enrich the company northbeam.io:
1. enrich_organization for industry, size, and funding
2. find_company_techstack for the tools they run
Return a short firmographic summary.enrich_organization returns industry, size, funding, and technologies from a domain. find_company_techstack returns the exact platforms the company runs.
Tech stack is an underused hook. Knowing a prospect runs a specific tool lets you open with a reason to care, not a generic pitch.
Step 5: Layer On Intent Signals
Data tells you who to reach. Signals tell you when. This is the layer that most enrichment tools skip.
Check buying signals for this contact and company:
- check_job_change (did they recently switch roles?)
- check_promotions (recently promoted?)
- company_job_openings (are they hiring for my ICP?)
- company_product_launch (any recent launches?)
Flag any that fired in the last 90 days.A new job, a promotion, a hiring surge, or a fresh product launch are all timing triggers. Reaching out the week a champion changes roles beats reaching out cold three months later.
Signal-based outreach reports 5–15% reply rates versus 1–2% for generic templates, per G2's sales intelligence benchmarks. Claude can attach the signal that fired as a note, so your outreach always has a "why now".
Step 6: Enrich a Whole CSV in One Prompt
Once the single-contact flow works, scaling it is a one-line change. Point Claude at a file instead of a name.
Read contacts.csv (columns: name, company).
For each row, enrich with SyncGTM:
- enrich_person for title + LinkedIn
- find_work_email, then verify_email
- enrich_organization for firmographics
- check_job_change for a timing signal
Write results to enriched.csv. Skip rows that fail.Claude loops the row, calls the tools per contact, and writes a clean, send-ready file. Start with 20 rows to check quality before you run 500.
This is the same pattern our Claude Code lead enrichment guide uses at scale — enrichment as a repeatable step, not a manual chore.
Tips and Best Practices
- Check credits before big runs. Call
check_creditsfirst. Mobile and phone lookups cost more per result than emails, so batch cost adds up fast on large lists. - Verify before you send. Always end an email pull with
verify_email. One bounce-heavy batch can damage your domain reputation for weeks. - Enrich narrow, then wide. Pull only the fields your outreach uses. You do not need a mobile number for an email-only sequence.
- Keep an ICP in CLAUDE.md. A shared context file lets Claude score and filter as it enriches, so you spend credits on fits, not noise.
- Let signals drive timing. Re-run
check_job_changeon open records to catch when a champion leaves before the deal goes quiet. - Human review stays. Claude drafts and enriches; you approve before anything sends. The review gate is where bad data gets caught.
Conclusion
Claude enrichment is not a new app to learn. It is the enrichment you already do, moved inside the tool where you build lists and write outreach.
One install, a plain-English prompt, and Claude pulls 76+ data points per contact — verified email, mobile, LinkedIn, tech stack, funding, and the signals that tell you when to reach out.
The model is the easy part. The data layer is what makes the output worth sending, and a waterfall across 50+ providers is what keeps fill rates high.
Start free on SyncGTM — no credit card required. Install the MCP in five minutes, and Claude enrichment has a real engine behind it from the first prompt.
