Company Spotlight: UpSmith’s AI Engine for Home Services

UpSmith is building AI for a part of the economy where software has often stopped at the dashboard. Home-services operators already have years of customer history, estimates, equipment records, memberships, and job notes. The problem is finding enough people and time to act on all of it.

The Dallas company is turning that unused operating history into work for AI agents. Its software confirms appointments, follows up on estimates, reactivates dormant customers, renews memberships, and identifies replacement opportunities inside the systems contractors already use. The near-term opportunity is more booked work. The larger question is whether AI can become a dependable capacity layer for businesses that cannot solve every bottleneck by hiring another customer-service representative.

What UpSmith Builds

UpSmith focuses on HVAC, plumbing, and electrical contractors. The company groups its product around customer experience, retention, and growth. Those categories translate into practical jobs: responding to inbound leads, handling appointment confirmations, following up on unsold estimates, recovering lapsed memberships, and finding old equipment or inactive accounts that merit another conversation.

The platform is designed to work behind the scenes rather than create another interface for office staff. UpSmith is a certified ServiceTitan Marketplace partner. Its integration reads customer, job, equipment, membership, and note data, then writes appointments and tags back into ServiceTitan so the dispatch board remains the operational record.

That integration is the center of the product thesis. A general-purpose model can write a message. A useful agent for a home-services operator also needs to understand why the customer is being contacted, what happened on the last job, whether an estimate remains open, when equipment is likely to need replacement, and when a human should take over.

The Business Case Is Hidden in Existing Data

UpSmith's argument starts with a constraint familiar to operators: the database contains more possible work than a customer-service team can pursue manually. Old estimates cool down, missed calls go unanswered, memberships lapse, and customers with aging equipment disappear from the active schedule.

The company says more than 200 home-services operators use its software. Its customer stories report that Five Star Home Services reduced cancellations by 34% while working across more than 180,000 customer records. Medley AC reportedly generated $147,000 in new revenue over three months and cut cancellations by 25% without adding headcount. Those are company-reported outcomes rather than independently audited benchmarks, but they show the commercial result UpSmith wants buyers to measure: booked work and recovered capacity, not chatbot engagement.

This is also why attribution matters. If an agent revives an estimate or prevents a cancellation, an operator needs to see the path from message to appointment to completed job. UpSmith's current product emphasizes revenue attribution by capability and timestamp. The quality of that measurement will matter as much as the fluency of the conversation.

Wyatt Smith Carried the Mission Into Software

Founder and CEO Wyatt Smith started UpSmith in 2022 after work spanning Teach for America, McKinsey, and Uber Elevate. A 2026 Harvard Business School profile describes how the company moved from workforce development toward AI software for the employers who create skilled-trades jobs.

The mechanism changed, but the underlying mission remained recognizable. Instead of training workers and matching them into roles, UpSmith is trying to help contractors grow the businesses that employ them. More booked work can create more productive hours for technicians and more room for operators to invest in their teams.

That connection keeps the company close to an important customer reality. A contractor does not buy AI because the category is fashionable. The buyer wants fewer empty dispatch slots, faster follow-up, more renewed memberships, and less manual work for the people already answering phones and coordinating jobs.

Distribution Runs Through the Trades Ecosystem

UpSmith's go-to-market advantage may be as operational as it is technical. ServiceTitan already anchors the workflows of many scaled contractors, and UpSmith's native integration reduces the cost of introducing a new system. The company also has a strategic partnership with the Air Conditioning Contractors of America, giving it another route into the community it serves.

The product still has to earn trust at every interaction. UpSmith says its agents identify themselves when directly asked, route conversations to humans when appropriate, and adapt to each shop's scripts, pricing, service area, and tone. Those controls are not a side issue. A mistaken promise or poorly handled reschedule can affect a real customer relationship and a technician's day.

The announced roadmap widens that responsibility. Alongside text-based workflows, UpSmith plans an inbound voice agent and a unified customer timeline. Voice raises the bar for latency, accuracy, escalation, and disclosure. A unified timeline could make agents more context-aware, but it also makes data quality and permissions more consequential.

New Capital Funds Engineering and Distribution

On October 5, 2026, UpSmith announced a $10 million Series A led by SemperVirens. The company says the money will expand engineering and distribution. Allison Baum Gates of SemperVirens and Rob Biederman of Asymmetric joined the board.

The hiring pattern fits that plan. UpSmith's current openings span engineering, customer outcomes, and commercial roles. Recent technical listings emphasize decision systems, trustworthy data, production migrations, and client-facing attribution. Commercial roles focus on translating ServiceTitan data into a business case for owner-operators.

That mix is revealing. UpSmith is not only training conversational models. It is building the data systems, operational support, and sales motion required to deploy agents inside businesses where the outcome appears on a dispatch board and a profit-and-loss statement.

What to Watch

UpSmith has a clear wedge, credible distribution partners, and customer examples tied to measurable operating results. The open questions are durability and repeatability. The company has not disclosed revenue, pricing, retention, margins, or independently audited outcome data. It still has to show that its agents perform reliably across different shops, seasons, customer histories, and service lines.

The strongest version of this company will not be judged by how human its messages sound. It will be judged by whether a contractor can recover more work without losing control of the customer experience.

Home-services businesses already possess the raw material: years of relationships and operational history. UpSmith is betting that the next important employee is a system that can finally put that history to work.