AI is here.
Let’s talk about it.
Used well, AI can save time and make someone far more productive than they were before. Used carelessly, it may as well be a fire in your office. Employees paste customer, financial, or personnel data into tools nobody has vetted, trust privacy controls they don’t fully understand, or accept polished output as correct. That information may then be stored or used in ways the business never intended.
I help businesses decide where AI is genuinely useful, where it isn’t, and what controls belong around it. When building bots and scripts for repetitive work, I can recreate real forms using synthetic information so the automation can be developed and tested without repeatedly exposing live business data. That reduces unnecessary data exposure as well as token and usage costs.
I also help employees understand the tools and use them with better judgment. The goal is to get the productivity without handing over the steering wheel, while making sure your people aren’t left scrambling to catch up later.
Where I help
From the first “should we even use this?” to the tools your team runs every day.
Where it fits
Where AI is genuinely useful in your business, and just as important, where it isn’t worth it yet.
Guardrails
The controls that belong around it: which tools are vetted, what data never goes in, and how output gets checked before anyone relies on it.
Safer automation
Bots and scripts for repetitive work, built and tested on synthetic copies of your real forms, so live business data isn’t exposed along the way and usage costs stay down.
Your people
Helping employees understand the tools and use them with better judgment, so nobody is left scrambling to catch up later.
From my own work
What this looks like in practice.
Checked to the cent
I built a company’s in-house operations platform with Claude Code as my pair programmer, setting the requirements and checking every module against real records. Before anyone relied on its invoice builder, it had to reproduce 300 real invoices to the cent. See File 01.
Someone else at the wheel
I coached a senior stakeholder to run Claude Code on their own for their website and automation projects. The tools stay useful after I step back because the people using them know how to drive.
Draft note“From my own work” is new. More ideas for this page: a short FAQ (“Is ChatGPT safe for our customer data?”) or what a starter AI policy covers. Delete this note when done.
Not sure where to start?
Tell me what your team is already using, or what you’re worried about. I’ll tell you honestly whether it needs attention.
Get in touch