What I do

← Back to Home

AI agents, built for one business at a time

Most of my clients come to me with the same story. Somebody on the team spends hours every week copying data between tools, sorting an inbox, or chasing the next step in a process. It’s work that has to happen, but it doesn’t need a person.

That’s where I come in. I build an agent around that one workflow. It uses the tools you already have, follows your rules, and hands anything unusual back to a human. It isn’t a chatbot you bolt on. It quietly does the job.

Things I’ve built agents for: qualifying and routing leads, triaging email, pulling data out of documents, first-line customer support, keeping a CRM clean, and scheduling.

The tools I use

I pick whatever is proven in production, not whatever is newest. The tools change every few months. How I build doesn’t: predictable workflows, skills I can test, and logs you can read.

I’m not tied to one model. Grok Bot, Gemini Argon, Meta Muse or a cheaper model for simple steps: I route each step based on the task, the cost and how private your data needs to be.

What you get when an agent goes live

• A rulebook, a plain-English file that says who the agent is, what it may do, and when it has to ask a human
• Logging and alerts, so you can always see what the agent did and why
• Guardrails: input checks, output filters, and a human sign-off for anything high-stakes
• Documentation for every skill the agent has
• A plan for changes, because your business will change and the agent needs to keep up

How we’d work together

It starts with a call. No pitch deck. We look at your workflows and figure out where an agent would actually help. If it makes sense, we scope a small pilot with a result we can measure, build it, and see how it does before talking about anything bigger.

A good first candidate is a process that’s high-volume, a bit tedious, and easy to get wrong when people are busy.