Hey there,
Matteo here from Kinetyca.
For years, our GTM stack was a row of browser tabs.
Sourcing lived in one tab, enrichment in another, copy in a third, sending in a fourth. Each tool was good at its own job. The real work was everything in between: pulling a list out of one tool, cleaning it, pushing it into the next one, and checking nothing broke on the way.
That in-between work used to eat most of our time. At the start of this year, we changed how it works, and the stack has run this way ever since. Here’s what it looks like.
What changed
Claude Code moved into the middle.
It reads and writes to every tool in the stack directly. So the stack stopped being a set of separate apps and started running as one loop. A signal comes in and becomes a sourced lead, then a verified email, then a queued message. Nobody carries it across four screens to get there.
None of the tools below are new, and you’ll probably recognise most of them. The change is the layer that sits between them.
The operating layer
This is the brain, and it has three parts.
Claude Code runs the system end to end: the sourcing logic, the scoring, the copy, the campaign builds and the reply handling.
GitHub versions every skill and workflow Claude Code runs, so the system gets tested like software. When something changes, we can see exactly what changed and roll it back if it made things worse.
Supabase is the lead bank and the memory. Every enriched lead lives there and Claude Code works from it, so a new campaign starts from the data we already own instead of from zero.
Who gets contacted
Three layers decide who we reach out to, and this is where the loop starts.
Data. Prospeo sources the person and verifies their email and mobile in one step. That one step replaced a waterfall of four or five tools. Apollo covers the company and contact database, and BlitzAPI handles programmatic sourcing when we need volume.
Verification. BounceBan settles catch-all domains without sending a single email. DeBounce cleans the list before anything goes out.
Signals. RB2B resolves website visitors down to the actual person. Apify scrapes LinkedIn and social activity for buying signals.
This is where having one system in the middle pays off. A signal from Apify or RB2B doesn’t sit in a dashboard waiting for someone to export it. Claude Code picks it up and moves it through sourcing and verification without anyone handing it over.
How we reach them, and what happens next
Smartlead and EmailBison run the email sending. HeyReach runs the LinkedIn outreach.
Attio holds the pipeline and the relationships. iClosed turns interest into a booked call.
Notion, Figma, Slack and ClickUp sit around all of it for content, design and day-to-day ops.
Why the middle matters more than the tools
It’s easy to read a list like this and think the tools are the advantage. They aren’t. Anyone can sign up for every one of them this afternoon.
The advantage is in the middle. Once one system can read and write to all of them, the hours that used to go into moving data between tabs go back into the work itself.
It’s the same system behind $18M+ in client revenue.
Where to start
You don’t need all of these. Different combinations fit different budgets and stages, and the job each tool does matters more than which brand does it.
Start with the brain: Claude Code, a GitHub repo for your workflows, and Supabase as the memory. Then add one tool per layer, and add the next one only when the volume asks for it.
Since you’re on the list, here’s something to get you started.
We put our 10 best Claude guides on one page. They run from first setup all the way to the Claude Code automations we use. Start with the beginner playbook at the top and work down.
Compilation of our best Claude guides for 2026
I’m curious where your stack is right now. Which layer still runs on someone moving data between tabs by hand? Reply to this email and let me know!
PS. If you want this Claude system built and run for you, book a call with us here
