Witnessed

The Story Behind Witnessed

I'm building Witnessed because I kept coming back to a simple problem: when people rent something, there's usually no trusted record of its condition at the moment they receive it.

That sounds like a small problem until something goes wrong.

A tenant can lose a security deposit over damage they believe was already there. A guest can be charged for something that existed before check-in. Someone renting a car can receive a damage claim days after returning it. The frustrating part is that these disputes often aren't really about who is telling the truth. They're about who documented the condition better.

I started thinking about what a better solution could look like.

People already have the hardware. Everyone carries a camera in their pocket. What they don't have is an easy way to turn those photos into a structured, objective record.

The existing options didn't feel right for individuals. Professional inspection tools are designed for property managers, inspectors, and enterprise teams. They're often expensive, complicated, or require someone to conduct the inspection. On the other end, a phone camera gives you a folder of pictures — but not much else.

So I started building Witnessed.

I wasn't starting from zero. I'd already built NexusSentry — an AI visual audit platform used by construction sites and property managers to turn camera footage into structured compliance reports. Witnessed is what happens when you take that same engine and make it accessible to individuals. The AI reasoning, comparison logic, and structured output format were already battle-tested in production before Witnessed generated its first report.

The idea was simple: take photos of the thing you're receiving or handing back, let AI understand what is actually visible, and turn those photos into a professional condition report.

Instead of just storing 30-40 images, Witnessed organizes observations by room or area, identifies visible conditions, provides confidence levels, compares before and after documentation, and creates a shareable PDF report that both parties can reference.

The interesting part wasn't just getting computer vision to identify objects or damage. The bigger challenge was figuring out what the AI should actually do with what it sees.

That's when the product direction became clearer to me.

I didn't want Witnessed to be another AI damage detector.

The real product is trusted documentation.

Sometimes the most valuable result is finding damage. Sometimes it's proving there wasn't any damage. An all-clear report can be just as important as a report containing problems.

That changed how I think about the product entirely. Witnessed isn't there to decide who is right in a dispute. It's there to create a shared record of what was visible at a particular moment — before there is a dispute.

I'm starting with rentals and handoffs because the problem is easy to understand: apartments, Airbnb properties, rental cars, equipment, and other things where condition matters at the beginning and end of someone's responsibility. The numbers are significant — 507 million Airbnb bookings annually, 30-40 million apartment moves per year in the US, 48 million rental car users. Nearly 30% of tenants lose part of their deposit to disputes that better documentation could have prevented. Every one of those transactions is a potential Witnessed report.

But I think the underlying technology can go much further.

The same engine that documents a rental apartment can document a boat before a charter, equipment before a job site handoff, a vehicle before a sale, a storage unit before move-in. The category isn't rental inspection software. It's trusted condition records for any physical asset that changes hands.

The bigger vision is a world where a camera becomes a kind of digital witness — documenting the condition of physical things and creating structured evidence that people can rely on later. The same way DocuSign made "did you sign it?" a settled question, Witnessed is building toward making "did you Witness it?" the standard question before any handoff.

I'm launching on Product Hunt this week and building in public from here. The product is live at witnessed.nexussentry.ai — free to try, no account needed, first report in under 60 seconds.

If you've ever lost money because you couldn't prove what something looked like before you received it, I'd love to hear your story. And if you're curious about the technical side — how the AI comparison engine works, how we handle confidence scoring and edge cases, or what building a consumer product on top of a proven enterprise vision platform looks like — ask in the comments.