One agent per client.
A human in the loop.
I run GTM work through Claude Code agents: an agent for each client, with others that enrich data, verify the work, report on progress, and check account health. The human decisions stay with us.
Follow the work.
Select a step to see what happens, what comes out of it, and where judgment matters.
Agent work
Which accounts deserve a closer look?
Each client has a Claude Code agent that keeps its goals, context, and account work together. I start with the market and the customers we are trying to reach.
- What leaves this step
- An account brief with source context.
Keep the evidence attached so the next person can inspect it.
The tools behind it.
- Markets & people
- Syft, Apollo, Clay, and ZoomInfo for target accounts, buyer contacts, and signals.
- Conversations
- LinkedIn and founder-voice email campaigns in Smartlead. People review the outreach in Slack.
- The source of truth
- HubSpot, Attio, and Salesforce connections. Coverage reports and pipeline claims checked against the CRM.
- Experiments & feedback
- Landing pages, LinkedIn and Meta audiences, conversion tracking, and product tickets grounded in customer problems.
Useful systems leave
a trail you can trust.
The human decision, the action, and the resulting record should stay connected.
See the work behind it