A home services company

A home services company in a major Canadian city, nearly four decades old, with tens of thousands of homeowners served and a reputation built on reviews. The owner still prices every job himself. He wanted an AI-powered operating system that gets smarter every month and makes the business less dependent on him, built by one partner under one roof.

The challenge: One owner pricing every job, five systems that did not talk

Every quote went through the owner: an on-site visit, dozens of photos, and a price from experience, with an AI chat helping on about a quarter of them. Work orders were rebuilt by hand from the quote and the photos, five to fifteen minutes each. The crew schedule was a colour-coded spreadsheet updated overnight by an offshore assistant, and a missed rain call could leave a painter idle for a morning.

The data existed: a decade of quotes in the CRM, the invoices behind them, tens of thousands of customer records, and a photo library for every job, spread across five systems that did not talk to each other. Lead follow-up ran on phone qualifiers chasing a speed-to-lead target. And a previous software build had taken most of a large budget and delivered nothing, so trust had to be earned quickly.

The solution: A demo first, one dataset, a quoting engine

A clickable demo of the platform before the contract, built from the owner's own brief, so the first conversation was about a product rather than a proposal.

Full data access in the first two weeks: the CRM, the quotes, and the photo library, merged into one truth file across the company's history, with a branded dashboard on top that reports the numbers the owner tracks every Friday.

A quoting engine built from the company's own history: scope in, comparable jobs found, price out. First from text, then from the photos with a handful of follow-up questions. Every version is scored against held-out jobs before the owner sees it.

The pricing rules that lived in the owner's head are being written down for the first time, job by job, so the engine prices the way he does. Weekly check-ins with the operations lead, and a roadmap that runs through the year: quoting, then work orders and scheduling, then marketing and hiring.

Every year [the company] should become easier to operate, more consistent, more profitable and progressively less dependent on me.

Owner, in his brief to North Group

The outcomes: One live dataset, a quoting engine under evaluation

A decade of scattered records is now one dataset, with a live view of the business the owner never had. The quoting engine is in evaluation against real jobs, with each version measured before it ships.

The owner's pricing logic is becoming a system the company owns, the first step toward quotes that do not need him in the room. Work orders, scheduling, and marketing follow on the roadmap.