For small businesses & non-profits · 🍁

Let me review your AI

My day job is building AI systems for institutions where a mistake is a legal event. Most weeks I watch small and mid-sized businesses get sold the opposite of that discipline: tools first, questions never. This page is a passion project, not a practice. Nothing here is for sale; it exists because I think smaller businesses deserve the same honest conversation.

The questions worth asking

These are the same questions I put to billion-dollar institutions, sized for a business that does not have an AI team. Ask them of any vendor, or of me:

  • What is our AI actually doing? An honest audit of the tools you have adopted (or your staff adopted without telling you): what data goes in, what comes back, what it costs, and whether anything checks the output.
  • Which model, at what cost per case? The eval discipline from my four-axis method, scaled down. On one enterprise task, a mid-tier model matched the premium one at roughly a thirteenth of the cost. Your workload deserves the same measurement before you sign anything.
  • Where does our data actually go? For Canadian businesses this matters more every year. There are strong models and patterns that keep client data inside your borders, and most vendors will not volunteer that conversation. I wrote about the pattern in your agents inherit your data residency.
  • What would zero-to-production look like? Not a chatbot on your website. One real workflow, the one that eats your team's hours, taken from manual to governed automation with a human still in charge of the decisions that matter.

Why I do this

Because the discipline scales down and almost nobody brings it down. Big institutions get evals, audit trails, and honest architecture reviews. Small businesses and non-profits get a subscription and a promise. Democratizing that gap is the Canadian spirit as I understand it: being based in Toronto, I care that Canadian organizations in particular get a fair version of this conversation, including the parts about funding programs and data residency that rarely come up in a vendor pitch.

The honest fine print

This is not a consultancy and there is no rate card. I have a day job I love. If you are a small or mid-sized business trying to make a sane AI decision and you want a second pair of eyes from someone who does this at the deep end, reach out on LinkedIn and tell me what you are wrestling with. If I can help in a conversation, I will. Views and advice are my own, not my employer's.

> esc