Click-Ins: Turning Vehicle Damage into Data
- Evan J. Cholfin
- 13 hours ago
- 3 min read
Beyond the Hype: How Photogrammetry, Precision Engineering, and Real-World Automation Are Quietly Transforming Insurance
By Evan J. Cholfin
When most people hear about technology transforming insurance, they tend to imagine some sweeping AI revolution. But my conversation with Dima Geyzersky, CTO and founder of Click-Ins, pointed to something much more grounded—and, in many ways, much more important.
What stood out to me in this episode of AI Speed was that Click-Ins is not approaching the market as an “AI-first” company in the way that phrase is usually thrown around today. Dima made clear that while AI plays a role, their real advantage comes from combining multiple engineering disciplines—including photogrammetry, 3D modeling, computer vision, and structured reasoning—to build a system that can be trusted in high-stakes environments.
That distinction matters.
Because in a space like automotive insurance, the challenge is not simply identifying whether something looks damaged. It is determining what is actually there, what is new, what is pre-existing, what can be documented reliably, and what can stand up under scrutiny. As Dima explained, recording an image is often treated as if it automatically means understanding a scene. But that is not the same thing at all. The real leap is turning a simple vehicle photo into usable, defensible data.
That is where this conversation gets especially interesting.
We talk about how Click-Ins helps insurers understand vehicle condition at underwriting, how the system compares damage over time, and how similar technology can be used across rentals, auctions, marketplaces, and other automotive workflows. What I found compelling is that this is not framed as replacement technology. It is about augmenting the people already doing the work—giving inspectors, adjusters, and operators a much stronger set of eyes and a much more consistent standard of documentation.
Another thread running through the episode is trust.
Dima is direct about the limits of hype-driven automation, especially in industries where accuracy, evidence, and accountability matter. He talks about reducing hallucinations, keeping a human in the loop, and building systems that rely not only on models, but on knowledge, geometry, ontology, and engineering rigor. That perspective makes this conversation especially relevant right now, when so many companies are rushing to present AI as a magic layer that can solve deeply complex workflows on its own.
We also get into something I think many operators will appreciate: the difference between an impressive demo and a scalable product.
In Click-Ins’ case, that means near real-time vehicle condition reports, automated documentation, synthetic training data instead of customer data, and a system designed to work in the real world—not just in a lab or pitch deck. It also means understanding the edge cases, adversarial scenarios, and operational risks that can make or break trust in a product like this.
If you are interested in where insurance technology is actually heading, how imaging and engineering can create measurable value, or what it really takes to automate complex visual workflows responsibly, I think you will find a lot in this episode.
Because this conversation is not really about hype.
It is about precision.
It is about trust.
And it is about what happens when a photograph stops being just a picture—and starts becoming data.