Edge processing
A small on-site compute device analyses video locally. Useful where connectivity is limited, or where keeping video on site is preferred.
Platform
How continuous risk intelligence is produced
No new infrastructure and no complicated rollout. The platform turns the cameras already on site into live alerts for the people who can act, and structured risk data for underwriting. Here is the process, end to end.
- Detect
Live risk signals from the site, starting with the CCTV already running
- Act
Alerts reach the people on site in time to step in
- Learn
Structured risk data flows back to underwriting
- Price
Cover, terms and renewal reflect the real risk
Then it repeats. Every cycle adds data, and pricing moves closer to reality.
- 01
Connect to existing CCTV
The platform works with nearly all existing cameras over standard network protocols. No rip and replace, and nothing intrusive to install.
- Most current camera estates are compatible
- Edge (on-site) or secure cloud processing
- Set up quickly, with insights flowing soon after
- 02
AI watches the live feed
Video is analysed in real time for the agreed risk events. Faces are automatically blurred, so the system reads behaviour and hazards, not identities.
- More than thirty configurable detection types
- Facial blurring applied by default
- Sensitivity tuned to each site
- 03
Alerts reach the site instantly
The moment a hazard appears, the right people on site are notified so they can step in before it becomes an incident.
- Real-time alerts to the business
- Routed to the people who can act
- Configurable per camera and zone
- 04
Risk data flows to underwriting
With the customer’s explicit consent, behavioural and incident data is shared with the insurer, enriching pricing, validation and renewal decisions.
- Shared only under clear policy terms and consent
- New behavioural data for underwriting
- Validate what was declared against site reality
- 05
Fewer claims, better loss ratios
Prevention compounds. Fewer incidents mean fewer and smaller claims, healthier loss ratios, and a safer site for everyone on it.
- Lower claim frequency and severity
- Healthier loss ratios over time
- Safer sites and fewer days lost to injury
Accuracy
Tuned to the site, not one-size-fits-all
Detection is configured for each environment, with parameters and sensitivity adjusted to match the operational context and the risk appetite of the people acting on alerts. The AI models are also continuously fine-tuned, so accuracy improves over time. The aim is simple: alerts that are relevant enough to be trusted and acted on.
The full range of what the platform can watch for is on Use cases.
Real-time alertsRouted to the right people on site
Configurable detectionPer camera and per use case
Structured risk dataFlowing back to underwriting
Deployment
Edge or cloud, whichever suits the site
The platform runs where it makes most sense for connectivity, cost and how the customer prefers their video to be handled.
Secure cloud
Video is processed through secure cloud infrastructure, with no additional on-site hardware to manage.
Privacy by design
Reads behaviour, not identities
Video is sensitive, so privacy is built in from the start. Faces are automatically blurred, customers keep ownership of their data, and it is shared with insurers only under clear policy terms and with explicit consent.
Specific data handling is agreed per deployment. The full picture, including where processing happens, retention and access control, is on Security and data. For how this website handles your data, see our privacy policy.
Facial blurring by defaultIdentities are obscured automatically
Customer owns the dataShared only with consent and clear terms
UK GDPR compliantData minimisation and DPIA support
See the full flow on a live site
A focused 30 minutes, not a slide deck: live detections, the risk data your underwriters would receive, and what a pilot could look like on part of your book. No obligation.