NOSO LABS Is Building an AI Co-Pilot for the HVAC Tech in Your Basement

The Macro: The Trades Have a Skills Crisis That Training Cannot Fix

The home services industry in the United States is worth over $600 billion a year. HVAC, plumbing, electrical, roofing, pest control. These are not optional services. When your furnace dies in January or your air conditioning fails in August, you are calling someone that day. The demand is enormous, the margins are healthy, and the customer acquisition flywheel, once established, is powerful. By most metrics, home services is a great business.

Except for one thing. The labor market is a disaster. There are roughly 400,000 unfilled positions in the skilled trades right now. The average age of an HVAC technician is climbing steadily. Training a new tech takes 18 to 24 months. And the real problem is not just technical skill. A great field technician needs three distinct competencies: the ability to diagnose problems accurately, the ability to recommend the right solution, and the ability to sell that solution to the homeowner on the spot. Finding someone who is excellent at all three is like finding a unicorn.

Most shops solve this with specialization and oversight. Senior techs handle the complex diagnostics. Sales training programs try to teach techs how to present options and close. Managers ride along on calls to coach. It works, sort of, but it does not scale. You cannot put a manager in every truck. And the moment a talented senior tech retires, all their diagnostic knowledge walks out the door with them.

The existing software in this space does not address the core problem. ServiceTitan, Housecall Pro, and Jobber are excellent at scheduling, dispatching, invoicing, and CRM. They run the business side. But they do not help the tech standing in a homeowner’s basement figure out why the compressor is short-cycling or how to present a $7,000 replacement option versus a $1,200 repair. The field service management market is well served. The field technician intelligence market barely exists.

The Micro: A 2x Founder and a ByteDance AI Researcher Go After the Trades

NOSO LABS was founded by Winston Chi and Alex Xi. Winston is a two-time entrepreneur. His previous company, Butter Technologies, was acquired by GrubMarket. He has been focused on vertical AI applications and specifically chose HVAC as the starting point for NOSO. Alex is a PhD with experience as a Principal Engineer leading code LLM initiatives, with prior roles at ByteDance AI and Meta AI. The combination of a repeat founder who understands go-to-market in messy vertical markets and a deep AI researcher who has built production LLM systems is exactly right for this problem.

They came through Y Combinator’s Summer 2025 batch as a three-person team in San Francisco. The product builds AI agents that support field technicians across all three competencies: diagnosis, solution design, and sales.

On the diagnostic side, the AI agent helps techs work through complex troubleshooting scenarios. Instead of relying solely on experience and intuition, the tech can describe symptoms and get guided diagnostic workflows. The system draws on manufacturer documentation, common failure patterns, and equipment-specific knowledge that would take a human tech years to accumulate. A second-year tech with the AI agent can work through problems that would normally require a 15-year veteran.

On the sales side, the agent helps techs present options to homeowners in a way that is informative, not pushy. It can generate repair-versus-replace comparisons, calculate energy savings projections, and present financing options. The homeowner gets a professional consultation. The tech does not have to fumble through mental math or memorize pricing tables. The company frames this as helping techs “sell 10x better,” which is aggressive language but directionally correct if the AI handles the presentation layer competently.

The paperwork piece is the third leg. Techs spend a meaningful chunk of their day filling out forms, writing up reports, and logging information into field service management software. The AI agent automates this, capturing job details during the call and generating the documentation automatically.

They are actively hiring a founding engineer at $120K to $200K with equity, which suggests the product is in early development but moving toward production deployments. The HVAC vertical is the beachhead, with a clear expansion path into plumbing, electrical, and other home services categories.

The competitive field is thin. XOi Technologies offers video-based remote support for techs but is focused on connecting field techs to remote experts, not on autonomous AI agents. Augmentir provides connected worker platforms for manufacturing, not home services. ServiceTitan has enormous distribution in the home services market but has not built technician-level AI. The gap is real.

The Verdict

I think NOSO LABS is building for a market that is about to become very hot. Every major home services company is struggling with the same problem: not enough experienced techs, too much knowledge trapped in the heads of senior employees, and no scalable way to bridge the gap. AI agents that can augment technician capabilities on the job site are the obvious solution, and the fact that nobody has built a credible one yet means NOSO has a window.

The founder combination is strong. Winston has been through the startup lifecycle before and knows how to sell into fragmented, relationship-driven markets. Alex has the technical depth to build AI systems that actually work in noisy, real-world environments. An HVAC basement is not a clean data center. The AI needs to handle imprecise descriptions, ambiguous symptoms, and interruptions from homeowners asking questions.

In 30 days I want to see pilot results from real service calls. Specifically, I want diagnostic accuracy. If the AI agent correctly identifies the root cause faster than the tech would have alone, the value is clear. If it sends techs down wrong paths, it becomes a liability.

In 60 days the question is revenue impact per truck. Home services operators think in terms of revenue per truck per day. If NOSO can demonstrate a measurable increase in average ticket size and close rate for techs using the AI agent, the ROI conversation becomes simple. A tool that adds $200 per day per truck in revenue can charge almost anything.

In 90 days I want to understand the data flywheel. Every service call generates diagnostic data, outcome data, and sales data. Over time, NOSO’s models should get better at diagnosing specific equipment failures and recommending the highest-converting repair options. If the system learns from every call across every customer, the competitive moat deepens with every deployment.

The trades are not going to train their way out of the skills gap. They are going to AI their way out of it. NOSO LABS is building the tool that makes that possible.