← Back to VeratrixAI
Platform guide

Start with a
well-scoped pilot.

A practical checklist for your operations, IT and security teams.

This is a scoping guide, not an API reference or a product compatibility matrix. Confirm supported models, protocols, modules and deployment options with the technical team.

1. Define the operational problem

Choose a small set of priority events, the areas involved, the responsible team and the action an alert should trigger. Agree what a useful outcome looks like, including when an alert should be dismissed.

2. Review the infrastructure

Reusing cameras does not necessarily eliminate new compute, licensing, gateways, installation or network work. Legacy or analogue systems may need a compatible recorder or stream conversion.

3. Agree response ownership

Specify the first reviewer, acknowledgement expectations, escalation path and the incident record. Select alert channels only after confirming their integration, provider setup, cost and information-sharing requirements.

4. Set acceptance criteria before testing

Detection quality

Define reviewed examples, expected events and the method for identifying misses.

False alerts

Record false alerts by scenario and their effect on operator workload.

Timing

Measure model processing, delivery and human response separately.

Operating conditions

Record stream quality, site conditions and degraded or unavailable feeds.

5. Review the decision to expand

Compare results with the agreed criteria. Document limitations, unresolved issues and required changes. Agree the next scope and commercial terms rather than assuming a pilot automatically becomes a full deployment.

What to share in your first enquiry

Send a non-sensitive summary of the site type, approximate camera count and priority problem. Credentials, live stream links and sensitive footage are not needed for an initial discussion.