*Expected for suitable environments and camera positions. Dedicated sensors remain the right choice where contractual accuracy above 99% is required.
Your CCTV records activity. It does not explain it.
You already have cameras across entrances, stores, concourses and public spaces. But most of that infrastructure produces footage, not footfall intelligence.
Cameras running. Questions unanswered.
Visitors enter, move between zones and leave. Your CCTV records every step, but you still cannot see which entrances perform best, where people dwell or how movement changes throughout the day.
Campaign impact stays unclear.
Did the campaign increase visits? Did customers reach the promoted area? Without consistent visitor data, marketing performance remains difficult to prove. Every budget decision carries more guesswork.
Staffing follows assumptions.
A quiet entrance and a crowded food court need different responses. Without live occupancy, queue and visitor-flow data, teams react late, schedules miss demand and service levels suffer.
Prove the fit before deployment.
Send a 10-minute clip from your existing camera.
We test detection under your real lighting and angles.
You see counting and anonymisation before committing.
Turn live video into useful numbers.
Count visitors at entrances and key boundaries.
Measure occupancy, dwell time and crowd density.
Compare movement patterns across zones and locations.
Put CCTV analytics into daily decisions.
Direct staff towards demand as conditions change.
Measure campaign impact through visitor behaviour.
Feed data into PFM dashboards or Power BI.
PFM combines Isarsoft Perception with an on-site AI compute unit to add analytics to compatible existing IP cameras. The problem is not a lack of cameras. It is a lack of intelligence from cameras that are already there.
Stop recording. Start understanding.
A camera captures activity. CCTV analytics turns that activity into operational intelligence.
Without analytics
Footage without context
With PFM
Intelligence that drives decisions
From entrances to entire estates.
The same data supports operations, marketing, asset management and executive reporting.
From spill to spotless. Without a single call.
Every spill creates a slip risk, a complaint and a cost. PFM CCTV AI can close the gap between detection and action by connecting existing cameras, access control and an autonomous cleaning robot in one automated workflow.
Spot
AI sees the hazard first.
PFM CCTV AI detects a change in floor condition through your existing camera network. The alert is created as soon as the spill appears, rather than waiting for the next patrol.
Dispatch
The cleaning robot is tasked.
PFM AI sends the nearest available autonomous cleaner directly to the affected area. No radio call. No manual task assignment. No member of staff sent into a busy concourse.
Access
Secured doors open automatically.
The workflow connects with your access-control system. Doors unlock when required, allowing the robot to reach the area without human assistance.
Clean
The result is verified and recorded.
The robot completes the clean. AI confirms the result and logs the task, giving facilities teams a record of the response and the completed job.
One detection-to-action loop.
This connected workflow supports faster hazard removal, more consistent cleaning standards and fewer manual call-outs. It also frees staff for customer-facing work while reducing the time a slip-and-trip risk remains on the floor.
The process has been demonstrated in a live proof of concept: AI CCTV detected a change in flooring conditions, instructed an autonomous cleaner to attend and used access control to provide entry, with no human in the loop. The autonomous cleaning component is powered by ICE Co-Botics Next Gen.
The right analytics layer for existing CCTV.
PFM takes a camera-agnostic approach. We assess the field of view, camera height, lighting and business objective first. A narrow entrance, wide concourse and multi-store estate may use the same platform, but each needs a different configuration.
Powered by Isarsoft video analytics
PFM delivers this capability in partnership with Isarsoft, a specialist in AI video analytics. Isarsoft Perception processes compatible CCTV feeds locally and converts them into anonymous data such as footfall, occupancy, dwell time, queues and movement patterns.
PFM connects that intelligence to Toolbox, Advantage and your wider reporting environment. You get proven video analytics technology, shaped around your locations and supported by PFM’s retail and footfall expertise.
Existing IP cameras
Use what you own
Compatible IP cameras using RTSP can feed the analytics layer without a wholesale camera replacement. Standard Full HD footage at 12–15 frames per second is sufficient for many people-counting applications. Camera position and angle still determine performance.
AI compute unit
Local video processing
An on-site NVIDIA GPU compute unit runs the analytics models close to the camera network. One unit can serve multiple camera feeds, depending on configuration, workload and required analytics.
Edge anonymisation
Privacy by design
Video is processed locally. Faces and number plates can be pixelated during analysis, while raw footage stays inside the customer environment. What leaves the building is metadata: counts, occupancy levels and timestamps, not identifiable images.
PFM analytics platform
One operational view
Footfall, occupancy, dwell time and crowd data can flow into PFM Toolbox and Advantage. Teams see trends across one location or compare performance across a multi-site estate.
Business integrations
Data where teams work
Analytics can connect through MQTT and business-intelligence environments such as Power BI. Start with footfall and occupancy, then add dwell time, movement analysis, staff exclusion and alerts as the use case develops.
The wider CCTV AI proposition can also support live occupancy management, real-time alerts, health and safety monitoring, autonomous facilities management and multi-site reporting.