SAFVR

01 / Pilot Methodology

How a 30-day SAFVR pilot actually runs.

A pilot is a 30-day, scoped evaluation on your existing CCTV. Setup in week one, detections live by week two, integrations wired up in week three, written results in week four. No commitment to continue. The methodology is designed to support an honest go / no-go decision at the end of the window.

02 / The 30-Day Window

Four weeks. Four checkpoints. One written outcome.

Week 1

Setup & camera onboarding

  • Kickoff with site safety lead and IT
  • Connect to existing CCTV streams (RTSP / ONVIF / NVR)
  • Edge appliance or GPU host provisioned on customer infrastructure
  • Detection zones, classes, and thresholds configured for your site

Week 2

Detections go live

  • AURA Adaptive Safety Engine begins live inference
  • Daily review with safety lead to tune false positives
  • Baseline incident & near-miss patterns documented
  • Initial DETECT loop output shared with stakeholders

Week 3

Workflow integration

  • ACT loop wired into your existing routing (email, Teams, Slack, EHS system)
  • IMPROVE loop micro-training assignments configured for repeat patterns
  • PREVENT loop dashboard set up for safety committee review
  • Role-based access provisioned for site, regional, and HSE team members

Week 4

Review & decision

  • Written pilot report: detections, response times, leading indicators surfaced
  • Side-by-side comparison vs. your existing baseline
  • Recommendations for full deployment scope
  • No-pressure decision meeting — extend, expand, or close out

03 / What's included

From SAFVR

  • Edge inference appliance (loaned for pilot duration)
  • Configuration of detection zones, classes, and thresholds for your site
  • Integration with one EHS or notification system
  • Daily monitoring and tuning during the pilot window
  • Written pilot report with documented results
  • Direct line to a SAFVR engineer throughout the pilot

04 / What you provide

From your team

  • Read-only access to existing CCTV feeds (RTSP / ONVIF / NVR / VMS)
  • Network path from cameras to the edge appliance (no cloud upload required)
  • A site safety contact and an IT contact for kickoff and weekly check-ins
  • Permission to deploy the edge appliance on your network

05 / Success Metrics

What we measure during the pilot.

Time-to-detection

How fast hazards are surfaced versus your current reporting baseline.

Repeat-pattern reduction

Whether IMPROVE loop micro-training closes the gap on recurring near-misses.

Leading-indicator coverage

Completeness of the picture: are unsafe acts visible before they become incidents?

06 / FAQ

Pilot questions, answered.

Does SAFVR send our footage to the cloud?
No. By default, raw video stays on your infrastructure. Inference runs on the edge appliance we provide (or your own GPU host). Only structured event metadata — timestamps, detection class, zone, and optionally redacted thumbnails — flows to the SAFVR control plane.
What does the pilot cost?
Pilot pricing depends on site size, camera count, and integration scope. Most 30-day pilots are scoped at a fixed engagement fee with no commitment to continue. Book a call to discuss scope.
What kinds of detections does AURA cover during a pilot?
AURA, our Adaptive Safety Engine, covers PPE compliance, restricted-zone entry, near-miss events, manual handling exposure, and customer-specific patterns. We confirm the detection set in the kickoff call before week 1.
Can the pilot run on existing cameras, or do we need new hardware?
SAFVR is designed to work with the cameras you already have. Most pilots run entirely on existing CCTV. If specific blind spots need coverage, we'll flag them — but adding hardware is never a requirement to start.
What happens if we don't continue after the pilot?
We retrieve the edge appliance, return or destroy your data per the DPA, and deliver the pilot report. No commitment, no auto-renewal, no penalty. The methodology is designed to support an honest go / no-go decision.
How is success measured?
We agree on success criteria in the kickoff — typically: time-to-detection vs. baseline, repeat-pattern reduction, completeness of the leading-indicator picture, and qualitative feedback from frontline supervisors. The pilot report compares actuals against those criteria.

07 / Start the conversation

Pick a 30-minute window. We'll scope your pilot together.

On the call: cameras, sites, integrations, success metrics, and a written scope you can take to your team. No prep required.