Turn the cameras you already own into events you can act on.

Questions buyers ask us

FAQ

What does Vision Intelligence add to cameras we already have?

Real ML detection and classification running on the existing camera feeds: people, vehicles, behaviors, anomalies. The cameras stay in place. The analytics layer reads frames over RTSP or a vendor API, runs inference on the edge or on a server, and pushes structured events into your VMS, alert channel, or access-control reaction. Older fleets work as long as resolution and framerate clear the basic threshold described above.

Do we have to replace our cameras?

Often no. Camera-agnostic overlay platforms run on existing fleets including five-to-seven-year-old systems. Vendor-tied analytics require their own cameras. Tec-Tel audits what you own and recommends replacement only where the existing hardware cannot carry the analytics you need. Phased upgrades across the contract are normal: keep what works, replace what does not.

How does the no-rip-and-replace pitch work technically?

Camera-agnostic platforms read existing feeds via ONVIF Profile S/T or direct RTSP, then run inference on a separate server or in the cloud. Vendor-specific platforms read via the vendor open API or the camera analytics chip if compatible. As long as your cameras output a usable stream and your VMS allows the integration, the analytics layer mounts on top without touching the cameras.

Will this lock us into a single AI vendor long-term?

Not unless you choose to be. Cloud-native platforms tie analytics to their cameras, so a switch is closer to rip-and-replace. Open platforms and camera-agnostic overlays let you switch the analytics layer without changing cameras. Tec-Tel designs for portability when the buyer flags it, which most multi-site operators do.

What's the realistic accuracy?

Vendors quote single-digit false-positive rates for object detection in tuned environments, with behavior analytics running higher. Real-world numbers start higher than the spec sheet and drop as zones, masks, and confidence thresholds get tuned. Plan for a tuning window and a pilot zone before site-wide rollout. The manual video-review scrub collapses to a filtered queue once the system is tuned.

What about employee monitoring law and BIPA?

Cameras can capture employee activity, but legal limits vary by state. Illinois BIPA requires written informed consent before collecting biometric identifiers like face geometry. Connecticut, Delaware, and New York have employer notice requirements. We default to badge or PIN credentials over biometrics unless the customer has a documented, consent-based program.

How much does adding AI typically cost?

AI analytics is a software line on top of fully-installed-camera cost, scaling with the tier of model and the number of detections licensed. Camera-agnostic overlays are priced per camera per month, which lets you scope the spend to the cameras that need analytics rather than the whole fleet. We itemize the analytics line on every proposal.

Which capabilities should we start with?

Not every site needs every capability. The free consultation picks three or four detections that change the operational picture for your environment (often PPE or restricted-zone, proximity, and intelligent search), proves them in a pilot zone, then expands once supervisors are actioning what they get.