Reconero links AI-written police reports to the footage behind them

Reconero wants police officers to be able to check an AI-written report against the recording behind it. The company says its software drafts narratives from body-camera footage and links each line to a video timestamp, leaving the officer to review and sign the result.

That connection between a sentence and its source is the most concrete part of a broader pitch: a single platform for the administrative work surrounding police cases. Founded by Riwan Azeroual and Edward “JiuJiu” Liu, Reconero combines video review, form preparation and searches across departmental policies. Its launch announcement describes software that works with existing records and dispatch systems rather than requiring departments to replace them.

The problem is easy to picture. One arrest can produce several documents that repeat the same facts in different formats. Reconero uses a DUI case as an example, with an incident report, crash report, consent form and toxicology request potentially following the same event. Its proposed workflow starts with the recording, identifies the relevant forms and prepares material for an officer to check.

Watch Reconero’s original launch video.

Video is also a search problem. An officer looking for a particular moment may otherwise have to scrub through a recording manually. Reconero says it processes body-camera, dashboard-camera, CCTV and interview footage, marking events such as searches or sobriety tests so users can jump to the relevant section. Those timestamps also provide a way back from a generated narrative to the underlying evidence.

For policy questions, the company describes a voice and text interface that returns answers with references to departmental rules and state law. The same platform offers plain-language queries over operational data, including charts of arrests or case-clearance figures. These are capabilities described by the company; the announcement does not provide an independent evaluation of their accuracy.

Azeroual previously worked in defense at Safran, while Liu studied computer science and engineering at Stanford, according to the launch. The founders say conversations with more than 100 officers helped shape the product. That figure describes interviews, not paying customers or departments using the software. The announcement does not disclose deployment counts, revenue or measured reductions in paperwork time.

The proposed integration model matters because the product is meant to sit alongside the systems departments already use. Reconero says it can read existing records-management and computer-aided dispatch systems without a migration. The launch does not spell out which vendors are supported or what an individual department’s installation requires.

For a department considering the product, the practical question is how much checking remains after the software finishes its draft. Linking a line to a recording makes review more navigable; it does not establish that the line accurately captures the event or its context. Reconero’s own workflow keeps that responsibility with the officer, who reviews the paperwork and signs it before it becomes a completed report.