The screen shows a driver’s view of a wet road at dusk, headlights reflecting off slick asphalt. A car ahead brakes, its taillights flaring red. For a human, it’s a moment of mild tension. For InsureVision.ai’s model, it’s a data point in a transformer network, a fractional uptick in a risk score that will quietly rank this driver against hundreds of others in a fleet manager’s dashboard tomorrow morning. The company’s bet is that this silent, contextual read of the road,what it calls “enviromatics”,is worth more than all the hard braking and sharp cornering data a traditional telematics box can provide.
Founded in London in 2022, InsureVision.ai is selling insurers and commercial fleets a new kind of predictive lens. Instead of relying solely on vehicle movement from accelerometers and GPS, its AI analyzes forward-facing dashcam video to assess the driving environment itself: traffic density, road conditions, and the behavior of other road users. The output is a cloud-delivered risk score, a ranking of drivers by their likelihood of generating a claim. It’s a product built on the premise that to see a crash before it happens, you need to see everything the driver sees, and understand it a little better.
The enviromatics wedge
Traditional telematics has long told fleet managers what a driver did,a hard stop here, rapid acceleration there. InsureVision aims to explain why it happened, and whether the context was actually dangerous. This shift from measuring driver inputs to interpreting environmental context is the core of its wedge into a market dominated by players like Nauto and Netradyne.
The company claims this approach yields a significant accuracy advantage. An independent review of two years of real-world testing, covering over 635 driver-years and 86 fault claims, found InsureVision’s video-based analysis delivered a 3x improvement in risk prediction accuracy compared to traditional telematics inputs [insurance-edge.net, Oct 2025]. The models are trained end-to-end on tens of thousands of real dangerous events and crashes, aiming to build a “fully learned crash prediction transformer” [InsurTech List]. For customers, the promised payoff is direct: data-backed fleets have negotiated 10-20% premium reductions, according to the company [Insurevision.ai].
A dual-path to market
InsureVision is approaching its market from two angles, reflected in products launched at CES 2026. For commercial fleets, there’s FleetVision, a video-based underwriting and risk assessment tool. For the emerging world of connected personal vehicles, there’s TeslaStick, a device that plugs into a Tesla’s glovebox USB port to access the continuous dashcam feed for personal lines insurance [insurance-edge.net, Jan 2026].
The company describes itself as “dash cam agnostic,” offering a cloud service that can ingest video from existing hardware, but it also manufactures its own AI dashcams [CB Insights]. This dual strategy,software-as-a-service and hardware,allows it to integrate with legacy fleet systems while also controlling the sensor stack for customers starting from scratch.
| Product | Target Customer | Core Function |
|---|---|---|
| FleetVision | Commercial Fleets & Insurers | Video-based risk scoring for underwriting & driver coaching |
| TeslaStick | Personal Vehicle Owners / Insurers | Access continuous Tesla dashcam feed for embedded insurance risk assessment |
| AI Dashcam | Fleets without existing video | Proprietary hardware with integrated vision processing |
Validation from a strategic corner
The $2.7 million seed round closed in March 2025 provides more than just capital. The investor list includes Rethink Ventures and Twin Path Ventures, but the most telling name is State Farm Ventures [InsurTech List, March 2025]. The participation of the venture arm of one of the largest auto insurers in the United States is a potent signal. It suggests not just belief in the technology, but a potential pathway to early adoption and rigorous, real-world validation within a massive insurance portfolio. The company says trials are underway with major insurers in the US and scheduled for Japan [InsurTech Insights].
The road ahead is crowded
The ambition is clear, but the lane InsureVision wants is already occupied. The competitive field includes well-funded, established players with deep fleet footprints.
- Nauto. A pioneer in AI-powered vehicle safety, backed by SoftBank and General Motors. It has a vast installed base and a strong focus on commercial fleet prevention.
- Netradyne. Known for its Driveri vision-based driver recognition and scoring platform, it has significant traction with large logistics and delivery companies.
- WingDriver. Focuses on smartphone-based driver coaching, offering a lower-friction, hardware-free entry point.
InsureVision’s answer to this competition rests on its specific technical claim of superior predictive accuracy and its “enviromatics” framing. Its ability to work alongside existing telematics, rather than demanding a rip-and-replace, is a practical wedge for sales conversations. The unanswered question is whether a 3x accuracy improvement, as compelling as it is in a study, translates into sufficient cost savings or loss prevention to compel a fleet manager to switch vendors or an insurer to redesign a underwriting model.
The solo founder question
Mark Miller, the founder and CEO, is steering the company through this competitive landscape. The public record shows a solo founder building both the AI technology and the commercial partnerships. While the seed funding and strategic investor validate the initial thesis, the next phase,scaling sales, managing hardware supply chains, and navigating enterprise procurement,will test the breadth of the team’s execution capacity. The company operates under the UK entity Advanced Automobile Solutions Ltd [Insurevision.ai].
What to watch in the next 12 months
The coming year will be about converting validation into volume. The milestones are straightforward, if hard to achieve.
- First major insurer deal. A publicly announced partnership or pilot with a top-20 auto insurer, beyond the current trials, would be a powerful traction signal.
- Fleet expansion. Moving from trials to deployed contracts with commercial fleets, demonstrating that the integration story works as smoothly as promised.
- The next funding round. With a seed round in place, the path likely leads to a Series A to fund scaling. The metrics that round is priced on will reveal much about the early market fit.
The product begins with a video feed and ends with a number in a spreadsheet, a reduction of chaos to a column sortable by risk. But the cultural question it’s built to answer is older than telematics: How do you measure carefulness? For decades, the industry has used proxies,speed, braking force, cornering g-forces. InsureVision’s bet is that carefulness isn’t just a function of the driver’s foot, but of their perception and reaction to a world of wet roads and sudden brake lights. It’s selling the idea that true risk hides not in the vehicle’s motion, but in the frame just outside the windshield.
Sources
- [InsurTech List, March 2025] InsureVision.ai company profile | https://insurtechlist.com/companies/insurevision/
- [InsurTech Insights] InsureVision Raises US$2.7 Million Seed Funding to Transform Vehicle Risk Assessment with AI | https://www.insurtechinsights.com/insurevision-raises-us2-7-million-seed-funding-to-transform-vehicle-risk-assessment-with-ai/
- [insurance-edge.net, Oct 2025] InsureVision using AI data to improve driver risk assessment | https://insurance-edge.net/2025/10/14/insurevision-using-ai-data-to-improve-driver-risk-assessment/
- [CB Insights] InsureVision.ai company profile | https://www.cbinsights.com/company/insurevisionai
- [insurance-edge.net, Jan 2026] InsureVision launches TeslaStick and FleetVision at CES 2026 | https://insurance-edge.net/2026/01/07/insurevision-launches-teslastick-and-fleetvision-at-ces-2026/
- [Insurevision.ai] Company website | https://insurevision.ai/