GTM engineering with Claude Code: the complete guide

GTM engineering is the practice of building go-to-market systems with code instead of stitching together SaaS tools with manual workflows. It is the reason a single operator can now do what used to require a team of five SDRs, a sales ops manager, and a data vendor contract.

The title barely existed two years ago. Now it shows up in job postings at companies like Rippling, Ramp, and Brex. LinkedIn reports a 400% increase in profiles with "GTM Engineer" since 2024. Something changed.

What changed is that AI coding agents got good enough to replace entire categories of manual GTM work. You no longer need a GUI to enrich data, find emails, or build prospect lists. You describe the workflow in natural language, and the agent executes it. Claude Code with gtmcli is the clearest example of this shift.

This guide covers everything: what GTM engineering is, why it matters, how it works with claude code, and how to get started today.

What is GTM engineering

GTM engineering sits at the intersection of sales operations, data engineering, and software development. A GTM engineer writes code (or instructs an AI agent to write code) that automates the repetitive work in a go-to-market motion: finding prospects, enriching data, validating contact info, personalizing outreach, and syncing everything to a CRM.

The key distinction from traditional sales ops: GTM engineers treat workflows as software. They version control their prospecting logic. They write tests for data pipelines. They deploy enrichment scripts that run on a schedule. When something breaks, they debug it like a production system, not by clicking around a SaaS dashboard.

This matters because sales ops has a compounding complexity problem. Every new tool adds another integration point, another API to manage, another monthly invoice. GTM engineering collapses that complexity into code that you own and control.

The GTM engineer skill set

You do not need to be a senior software engineer. Most GTM engineers have a background in sales ops, revenue ops, or growth marketing. They picked up scripting because they got tired of doing the same thing manually every week. The typical GTM engineer can:

Write basic Python or JavaScript. Use APIs to pull and push data. Think about data quality systematically. Prompt an AI agent to build what they cannot code themselves. That last point is the one that changed the game. You no longer need to write every line yourself.

The old workflow: why it stopped working

Here is how most B2B companies built prospect lists before GTM engineering became a discipline:

The old workflow (per 100 contacts):

1. SDR searches LinkedIn Sales Navigator     — 45 min
2. Export to spreadsheet, clean up data       — 20 min
3. Paste domains into Clay or similar tool    — 10 min
4. Wait for enrichment to run                 — 15 min
5. Copy enriched data back to spreadsheet     — 10 min
6. Run emails through a validation tool       — 10 min
7. Remove bounced/invalid emails              — 5 min
8. Format for CRM import                      — 10 min
9. Upload to CRM, create sequences            — 15 min

Total: ~2.5 hours per 100 contacts
Monthly cost: $500-2000 in tool subscriptions

Every step involves a different tool. Every tool has its own login, billing, and failure modes. Data gets copy-pasted between tabs. Columns get misaligned. Emails expire between the time you find them and the time you send. The whole process is fragile, slow, and expensive.

Some teams tried to fix this with Clay or similar workflow tools. Those helped with the orchestration, but you still had to build the workflow in a GUI, pay per-credit for every enrichment, and deal with a visual interface that could not be version-controlled, tested, or shared as code.

The new workflow: claude code + gtmcli

Here is the same task with claude code and gtmcli:

The new workflow (per 100 contacts):

1. Open claude code                           — 0 min
2. Describe ICP and target list               — 1 min
3. Agent researches, finds, validates          — 4 min
4. Review CSV output                          — 2 min
5. Agent pushes to CRM via API                — 1 min

Total: ~8 minutes per 100 contacts
Monthly cost: Pay per result, typically $5-15

That is a 95% reduction in time and an 80%+ reduction in cost. But the biggest gain is not speed or money. It is repeatability. The agent remembers your ICP. You can rerun the same workflow next week with a single prompt. You can modify it incrementally. You can share the prompt with your team.

How claude code and gtmcli work together

Claude code is an AI coding agent that runs in your terminal. It can read files, write code, run shell commands, and call external tools through MCP (Model Context Protocol). It is not a chatbot. It is an agent that takes actions.

gtmcli is a command-line tool and MCP server for go-to-market data. It provides email finding, email validation, and company enrichment through a simple API. When connected as an MCP server, claude code can call gtmcli tools directly during a conversation.

The combination works like this: you describe what you need in plain English. Claude code figures out the plan. It calls gtmcli tools to get data. It processes, filters, and formats the results. It writes the output to a file or pushes it to an API. You review and approve.

You (prompt)
  → Claude Code (plan + execute)
    → gtmcli MCP tools (data)
      → find_email
      → validate_email
      → enrich_company
    → File system (write CSV)
    → CRM API (push contacts)
  → You (review output)

Setting up the stack

The setup takes five minutes. Install gtmcli, authenticate, and add the MCP config.

# Install
npm install -g gtmcli

# Authenticate
gtmcli auth login

# Add to ~/.claude/settings.json
{
  "mcpServers": {
    "gtmcli": {
      "command": "gtmcli",
      "args": ["mcp", "serve"]
    }
  }
}

Restart claude code. The agent now has access to email finding, validation, and enrichment tools. You are ready to start building.

Five GTM engineering workflows you can build today

1. Account-based lead lists

Give the agent a list of target accounts and persona criteria. It finds the right people, gets their emails, validates them, and exports a CRM-ready CSV. This replaces the LinkedIn Sales Navigator + email finder + validation tool chain.

> Here are 50 target accounts (see accounts.csv).
> For each, find the VP of Engineering and CTO.
> Get verified emails. Export to abm-list.csv.

2. Inbound lead enrichment

When a lead fills out a form, you often only get a name and email. The agent can enrich the company domain, pull firmographic data, validate the email, and score the lead based on your ICP criteria. Wire this into a webhook and it runs automatically.

3. CRM hygiene

Export your CRM contacts to CSV. Ask the agent to validate every email, flag bounced addresses, and find replacement emails for invalid ones. Import the cleaned list back. Run this monthly and your bounce rate stays below 1%.

4. Competitor customer research

Tell the agent to research companies using a competitor product. It can identify customers from case studies, review sites, job postings, and integrations pages. Then it finds the decision makers and their emails. This used to be a full-time research job.

5. Event-based prospecting

Give the agent a conference attendee list or a webinar registration list. It enriches each person with company data, finds their direct email (not the one they registered with), and builds a follow-up list segmented by ICP fit.

Before and after: a real comparison

We tracked a GTM team that switched from the traditional stack to claude code + gtmcli. Here are the numbers after 90 days:

Metric                    Before        After
─────────────────────────────────────────────────
Contacts sourced/week     200           1,400
Time per 100 contacts     2.5 hours     8 minutes
Email validation rate     72%           96%
Bounce rate               4.8%          0.6%
Monthly tool spend        $2,100        $180
Headcount (sourcing)      2 SDRs        1 GTM engineer

The team went from 200 contacts per week to 1,400. Not because they worked harder, but because the agent does the tedious work and the human focuses on strategy: which accounts to target, what messaging to use, when to follow up.

Building a career in GTM engineering

GTM engineering is one of the fastest-growing roles in B2B. Companies are hiring because they realize one person with the right tools can outperform an entire team using the old playbook.

If you are in sales ops or revenue ops today, learn to use an AI coding agent. Start with claude code. Connect gtmcli. Automate one workflow that currently takes you hours. Show the results to your team. That is the fastest path from ops to engineering.

If you are already an engineer interested in GTM, the opportunity is enormous. Most sales teams are still doing things manually. They need someone who can build systems, not just run them. The pay reflects the scarcity: GTM engineer salaries in 2026 range from $140K to $220K base, often with equity.

Getting started in 15 minutes

Here is the fastest way to see GTM engineering in action:

# 1. Install gtmcli (2 min)
npm install -g gtmcli
gtmcli auth login

# 2. Connect to claude code (1 min)
# Add MCP config to ~/.claude/settings.json

# 3. Build your first list (8 min)
# Open claude code and prompt:
> Find 10 VPs of Marketing at Series B
> fintech companies with 100-300 employees.
> Verified emails only. Export to CSV.

# 4. Review the output (2 min)
# Open the CSV. Check the data quality.
# You just built a lead list with code.

That first list is your proof of concept. Once you see it work, you will find a dozen other workflows to automate. CRM cleaning. Inbound enrichment. Conference attendee research. Domain-based company lookups. Each one is a prompt away.

GTM engineering is not about replacing people. It is about giving one person the tools to do what used to require a team. Claude code is the agent. gtmcli is the data layer. Together, they are the foundation of the code-first GTM stack.