AI Firm Offers Free Home Repairs — If Workers Can Record the Jobs to Train Robots

An AI firm is offering qualifying homeowners free home repairs in exchange for allowing technicians to record their work to help train robots.

Launched on 6 October 2026 by AI data and training company HUMXN, the programme covers selected plumbing, electrical, heating and air-conditioning (HVAC), air-duct cleaning and carpet-cleaning services in Minneapolis, Chicago and Miami.

Participating tradespeople use wearable recording equipment to capture their work from a first-person perspective, creating footage intended to help train artificial intelligence and robotics systems.

The offer is not available for every job.

Homeowners must agree to the recording before a visit is arranged, and HUMXN assesses each request individually. Large renovations, new construction, roofing and emergency calls are excluded. Unusually expensive parts or work beyond the agreed scope may also fall outside the coverage, with HUMXN saying it will discuss these costs with homeowners beforehand.

Which Home Repairs Qualify for the Free Service?

HUMXN's programme operates through participating service companies, with homeowners required to submit details of their proposed jobs for eligibility checks.

The company says qualifying work is generally limited to a normal service call or approximately a day's work. For approved appointments, HUMXN pays the participating service company, allowing homeowners to receive the agreed service without paying for the visit or labour themselves.

The programme covers selected plumbing, electrical, HVAC, air-duct cleaning and carpet-cleaning jobs. Residents in the three launch cities can submit a request, but a free appointment is not guaranteed: HUMXN says a member of its team reviews each proposal and confirms whether it qualifies.

Homeowners should clarify the agreed scope before work begins, particularly if a repair requires expensive replacement parts or additional work.

How AI Records Work Inside Your Home

During participating visits, tradespeople wear body-mounted recording equipment that captures tasks from their own perspective.

The footage can show a plumber investigating a leak, an electrician carrying out diagnostic checks or an HVAC technician examining a faulty system. It can also document the sequence of actions involved in completing a job, including tool selection, troubleshooting, adjustments and final testing.

HUMXN says its recording process focuses on technicians' hands, tools and the equipment being serviced. The company also says identifying details captured incidentally, including faces, names, mail and information displayed on screens, are blurred or removed before the footage is shared.

It says homeowners' addresses are not disclosed in the resulting datasets and that approved AI and robotics organisations can use the footage under licence to train and evaluate their systems.

These are HUMXN's stated privacy measures, rather than independently verified findings.

The company invites homeowners to ask questions about what will be recorded and who will have access before a visit is arranged. Residents should review the programme's terms and privacy information before agreeing to participate.

How the Footage Could Help Train Robots

The initiative targets a challenge in physical artificial intelligence: helping machines learn to interact with the real world rather than simply process information on a screen.

Robots designed to perform practical tasks need examples of how objects are handled, tools are used and workers respond when a job does not proceed as expected.

A written manual might explain how to repair a leaking pipe, for example, but a recording of an experienced plumber at work can provide additional demonstrations of the physical steps involved, from inspecting the problem to selecting tools and checking the result.

Footage collected in real homes could also expose AI systems to different equipment, layouts, working conditions and repair techniques. Such variation may help developers prepare systems for situations that differ from controlled demonstrations.

However, recordings capture observable actions rather than a worker's complete internal reasoning.

Nor does collecting footage mean robots can already carry out plumbing, electrical or HVAC repairs independently. HUMXN's stated aim is to develop training data for AI and robotics systems intended to understand and eventually perform physical-world tasks.

Why HUMXN Is Paying for Home Repairs

The programme creates a commercial arrangement involving homeowners, service contractors and AI developers. Residents can receive qualifying services without paying the agreed costs themselves, contractors are paid for their work, and HUMXN obtains footage that can be developed into training datasets.

'The next generation of AI is going to need more than text and images from the internet,' said Thomas Unise, founder of HUMXN.

'Robots need to understand how skilled people interact with the physical world. Home-service technicians perform millions of highly technical tasks every day. We are creating a way for homeowners and service companies to benefit financially from the value of that expertise.'

The model reflects an effort to turn practical skills into data that AI developers can use.

Instead of relying solely on written instructions or demonstrations created in controlled environments, companies can obtain examples of professionals handling real equipment and responding to everyday problems.

The usefulness of the resulting datasets will depend on factors such as recording quality, the range of tasks captured and their suitability for the systems being developed.

HUMXN's announcement describes the intended benefits of the programme, but does not establish that the footage has already improved robots' performance.

Why the Programme Is Launching in Three US Cities

HUMXN is starting with participating contractors in Minneapolis, Chicago and Miami, with plans to expand into additional regions and skilled-service categories.

The company says the geographic spread is deliberate, exposing its data collection to different climates, building types, equipment and service conditions.

An HVAC technician working in Miami, for example, may encounter different systems and problems from one working in Minneapolis. Plumbing work in an older Chicago property may also differ from a job in a Florida home.

For developers training physical AI systems, recordings from varied environments could provide a broader range of examples than footage collected from a single location.

HUMXN's Major Ambitions for Physical AI

HUMXN works with companies, experts and operators to source and develop specialised datasets for AI training and evaluation. Its home-service programme extends that approach into physical AI and robotics, using demonstrations of skilled work as a potential source of training material.

The company says the approach could serve developers working across industries such as construction, manufacturing, logistics and automotive repair, where AI systems may need to learn from real-world tasks.

'There are millions of skilled workers performing incredibly complex tasks every day, but very little of that expertise exists in a format that AI systems can learn from,' Unise said.

'We want to build the infrastructure that connects that knowledge with the companies building the next generation of intelligent machines while making sure the people creating the value participate in the economics.'

HUMXN expects demand for real-world task data to increase as developers work on systems capable of operating beyond controlled environments. That is the company's forecast, rather than a demonstrated outcome of this programme.

For homeowners, the proposition is straightforward: HUMXN may pay for an eligible home-service job if the homeowner agrees to recording and the work meets the programme's requirements.

For AI developers, the initiative is an attempt to turn demonstrations of skilled human labour into data that could help train machines for tasks in the physical world.