A priced quote in 45 seconds, for eight brands

In home services, the first company to give a price usually wins the job. But pricing a tree job normally means sending someone out to look at it, and by then the homeowner has called three competitors.

I engineered and run an AI estimator that prices a job from a single photo. It's deployed across eight tree-service brands, each under its own name. A homeowner uploads a picture on the website and gets a branded quote by email and text in about 45 seconds, while the lead drops into the company's CRM.

For people who would rather call, a voice agent answers, asks them to text a photo, and replies with a price. And because no estimator is perfect on day one, I built a loop that compares the AI's quotes with what each job actually sold for, company by company, so the pricing learns.

Instant quotes on the web and on the phone, no lead ever left without an answer, and an estimator built to get closer to each company's real prices with every job it sells.

GPT-5 vision, voice agents, Twilio, CRM integration, Supabase, n8n orchestration.