“Everything depends on someone doing it by hand.”

The manual path mapped as it really runs, not how the org chart says it runs, then rebuilt so it runs itself. I start by watching the actual steps, handoffs, and exceptions before automating anything. What you get back is a process that no longer depends on one person remembering to chase it.

• Make
• Zapier

Leads scored, enriched, and routed the moment they arrive, instead of sitting in a queue until someone gets to them. Missing firmographic or behavioural data is filled in automatically so the right owner picks up a complete record. Speed to lead goes up without adding headcount.

• HubSpot

Agent systems that share context and take real actions, not isolated chatbots answering one prompt at a time. Built in Cursor and wired into the tools the work lives in, with persistent memory so each run builds on the last. They operate on a schedule and hand off to a person when judgment is needed, rather than waiting to be asked.

• Cursor
• OpenAI
• Vercel

The weekly report that writes and sends itself, pulled straight from the clean data rather than a manual copy-paste. I build it on Node.js and GCP so it runs on a schedule and lands in the channel your team already reads. The morning that used to disappear into reporting comes back.

• Node.js
• GCP

Monitoring that makes silent breakages loud before they cost you, rather than after a customer notices. Syncs, pipelines, and key metrics are watched, with alerts routed to the people who can actually fix them. Small failures get caught while they are still small.

• Node.js
• GCP

A documented system your team can run and extend without me in the loop, because lock-in is not the goal. I write down how it works, why it is built that way, and what to do when something changes. You own the automation outright when the engagement ends.