Toronto, Canada

I lead AI at Jobber.
I also run my own life on it.

I build software that does the work instead of just organizing it. Right now that means an AI teammate for the people who run home service businesses: the plumbers, landscapers, and cleaners who did not get into the trade to do data entry.

What I believe

Nobody wants a dashboard. A contractor finishing his eighth job of the day does not want a better view of his unsent quotes. He wants the quote sent. Most software still stops one step short of the actual job.
The unit of AI product is the finished task, not the answer. Chat interfaces made AI feel useful. Completed work makes it valuable. The gap between those two things is where the next decade of product gets built.
Memory is the moat, not the model. Models commoditize on a quarterly cadence. What compounds is a system that knows how this specific business runs and gets sharper every week it is used.
Trust is earned in small reversible steps. People delegate work long before they delegate control. Draft it, show it, let them approve it. Earn the right to act on your own.

What I work on

I own AI product at Jobber. The work comes down to three questions I find genuinely hard.

Getting from advice to action

Software has become very good at telling people what they should do. The harder and more valuable problem is doing it for them, and earning enough trust to be allowed to.

One system instead of a pile of features

AI has made it cheap to add capability and expensive to stay coherent. Most companies are solving the first problem. I care more about the second.

Products that improve with use

Anything as good on day two hundred as it was on day one is a tool. I am interested in the other kind, and in what has to be true underneath for that to work.

Before Jobber I led AI workflows, payments, and invoicing product at Intuit for QuickBooks.

What I am building for myself

I do not think you can credibly build AI products without living inside them. So I run my own work on a system I built and keep rebuilding.

A chief of staff that reads before I wake up

Every morning it sweeps my email, calendar, and team channels and hands me a briefing: what moved, what is at risk, what needs me specifically. I read one page instead of four inboxes.

A memory layer that outlives the conversation

Decisions, context, and the reasoning behind them get written down automatically as I work. Six months later it still knows why we chose what we chose. This turned out to be the piece that mattered most, and the piece I underestimated.

Drafts waiting instead of a queue to clear

Replies, updates, and prep documents show up already written in my voice. My job is to judge and send, not to start from a blank page. It is the same draft-and-approve pattern I am betting on at work, which is exactly why I trust it.

Off the clock

I am training for a half Ironman, learning to cook meat over fire properly, and getting ready to meet my daughter this fall. My wife Connie is a founder, which means we talk about product at dinner more than either of us intended.

Get in touch

Building something in this space, hiring, or just want to argue about where AI products are heading? Send me a note.