How I work
Some businesses need one focused project.
Others need a systems thinker in the room as they scale.
Same three phases.
The map
01
Investigation
02
Assessment
03
Build
01
2 to 3 weeks. I listen.
I sit in your stack. I talk to your team, not just leadership. I look at how the work actually moves: where it gets stuck, where it disappears, who knows what only because they were there when it was decided. By the end you have my read on what’s broken, in writing.
02
A written report you actually use.
You get a written diagnosis: how your team is actually using your tools, what’s redundant, what can be automated, what needs to change, and in what order. This is yours regardless of whether we keep working together. For some businesses, the assessment is the whole engagement. For others, it’s the foundation we build from.
03
The system goes in. Your team learns it. You keep it.
I architect and build the system, train your team on it, and document it so you don’t lose it when someone leaves. Sometimes this is a focused 3-week project. Sometimes it’s deeper work over months. Either way, the goal is the same: you don’t need me forever.
A specialty
AI can’t work on broken systems.
Most teams try to layer AI on top of their existing chaos and wonder why it doesn’t deliver. AI inherits whatever’s underneath it. Bad data in, bad answers out. Scattered SOPs in, hallucinated SOPs out. Generic chatbot in, generic chatbot forever.
I do something different. I build custom Claude and GPT agents that integrate with your real tools, train on your real workflows, and respect how your team actually operates. Onboarding your AI is the same job as onboarding your team. Both require a working system to plug into. Both require ownership, training, and documentation. Both pay off compoundingly once they’re tuned in.
In a project, that might look like a custom GPT for the leadership team that drafts status updates from live Asana data. In an ongoing partnership, it might be a Claude agent that lives inside a Notion workspace, trained on years of SOPs, that new hires query instead of pinging a senior. In a training engagement, it might be teaching three super users how to build their own tooling on top of Claude or ChatGPT for the work they do every day.
The deliverable is not magic. It’s leverage. Built on a foundation that actually holds.
“Do I have to commit to all three phases?”
No. Some clients hire me for the assessment and run with the report themselves. Others need the full build. The investigation phase tells us what makes sense for your team.
“How involved does my team need to be?”
More involved than you’d think during investigation, less involved than you’d fear during build. I do the heavy lifting. Your team’s job is to be honest about how the work actually happens.
“What happens after the build is done?”
You have a system your team owns. I document everything I built. If something needs tuning later, I’m available, but you’re not locked in. The goal is a system you keep, not a dependency on me.
The best way to know what you need is to talk it out.