Plate recognition that works on real roads
Reads Kenyan plates in glare, at night and at angle — and infers direction from plate colour (white in, yellow out) at your entrances.
AutoVision is AI-driven fuel intelligence. It turns camera video into a live, measurable picture of every vehicle, pump and customer — from the main road to the till.
The problem
How many cars passed — and how many actually stopped?
AutoVision counts road traffic and compares it with unique vehicles entering the station.Which pump is the bottleneck, and which one sits idle?
Occupancy, dwell time and queue waits per pump — against that pump’s own normal.Of the drivers who fuelled, how many walked into the shop?
Pump sessions linked to doorway visits, with confidence bands instead of guesses.See which pumps are occupied, how long each vehicle has been there, and how that compares with the pump’s normal — the moment it happens.
Fig. 01 — Pumps performance, live view

Understand where people gather, when queues form and how busy each zone is across the day — without identifying anyone.
Fig. 02 — People Intelligence

The number every fuel retailer wants: of the vehicles that fuelled, how many became shop customers — measured, not estimated.
Fig. 03 — Shop Footfall & Pump → Shop

Link pump sessions to shop visits and see which pumps, hours and moments turn a fill-up into a basket.
Fig. 04 — Pump → Shop analytics

Executive summaries, period comparisons and operational insights — written in plain language and generated from your live data.
Fig. 05 — A real Executive Daily Report

The platform
AutoVision replaces five disconnected tools with a single AI-driven fuel intelligence platform — every vehicle, pump, person and shop visit in one connected picture.
Capabilities
One platform instead of five disconnected tools.
Reads Kenyan plates in glare, at night and at angle — and infers direction from plate colour (white in, yellow out) at your entrances.
Evidence-backed recommendations, a “what if” revenue simulator, and action tracking that measures before and after — so you learn what actually worked.
Crowd build-up, rapid departures, restricted-zone entry, loitering, lighting failure — plus service moments like bonnet-open and windscreen cleaning.
Blacklist, whitelist and stakeout patterns. A match triggers email and webhook notifications automatically.
Find the vehicles that keep coming back, how often, and through which entrance — with a sensible merge window so quick re-reads don’t inflate loyalty.
Drag pumps, cameras, the shop and exits into place to mirror your real site — then watch vehicles and occupancy move across it, live.
Invite-only access, roles from viewer to super-admin, per-company data isolation and a privacy mode for non-admin operators. Ask the built-in assistant if you get stuck.
How it works
Cameras placed for each job — entrances, forecourt, shop door, main road — with zoom cameras where the view demands it.
A GPU-accelerated AI pipeline reads plates and tracks vehicles and people in real time.
Events are stitched into sessions: road, entrance, pump, shop, exit — each with evidence.
Dashboards, alerts, reports and recommendations your team can use the same day.
Why AutoVision
Fuel retail runs on thin margins and fast decisions. AutoVision replaces opinion with evidence.
Cameras record everything and explain nothing. AutoVision turns footage into counts, durations and conversions you can act on.
Every metric traces back to camera events and photos, and the product states plainly what a number is — and what it isn’t.
Spot idle pumps, long queues, drive-offs and low pump-to-shop conversion — then test changes and measure the result.
Local plate formats, motorcycles, trucks and buses, glare and night scenes, Nairobi time — tuned on a live Kenyan forecourt.
Every forecourt is different. AutoVision is designed per site — including zoom cameras where a lane or pump sits far from the lens — so the AI gets a clear view of what matters.
Invite-only access, role-based permissions, per-company isolation, audit logs, and people analytics that never identify individuals.
Where it fits
Pump analytics, shop conversion, queues, repeat customers and safety signals.
Count passing traffic and measure how much of it your site captures.
Dwell time, policy enforcement and watchlist alerts at every gate.
Tag fleet plates, follow journeys across cameras, integrate by webhook.
Blacklist alerts, stakeout detection and audit-ready logs.
One platform, many stations, with scoped access for every team.
FAQ
AutoVision is AI-driven fuel intelligence for Kenyan petrol retail, not just an ANPR camera app. It unifies forecourt analytics, pump occupancy, shop footfall, live vehicle sessions, queue and safety monitoring, hotlist alerts and executive reports in one platform.
No. ANPR/LPR is one module. AutoVision correlates plates with pump zone occupancy, shop visits, dwell time, main-road traffic counts, period-over-period charts, SMS and email alerts, and multi-camera RTSP roles (entry, exit, forecourt, shop) on a single operator dashboard.
Yes. Operators draw custom pump zone IDs on live camera preview (diesel, petrol, LPG, bay labels). Entry cameras open a vehicle session; pump and shop events attach automatically; exit closes the session with dwell and fueling duration.
Yes. AutoVision includes NTSA-tuned plate recognition with OCR correction for common misreads. Plates are validated, tracked across cameras, and searchable — as part of the broader fuel station intelligence workflow, not as a standalone LPR tool.
Hotlist and stakeout plate matches trigger SMS (TextSMS Kenya), email, Slack, and webhook notifications. Security and forecourt teams get real-time alerts when flagged vehicles enter or move across the site.
Fuel stations, petrol retailers, forecourt operators, and security teams across Kenya that need unified site intelligence — sessions, pumps, shop, traffic, and alerts — with ANPR included where it matters.
AutoVision is specified for each site. Entrances, the forecourt, the shop door and the main road are covered with cameras placed for the job, and zoom cameras are added where a lane or pump sits far from the lens, so the AI gets a clear view of what matters.
Use the contact form on autovision.co.ke, email admin@autovision.co.ke, or call +254 736 326 392. AutoVision is based at Nairobi Pentecoastal Bible College, Suite 23, Garden Estate Rd, Thome, Nairobi, Kenya.
Get AutoVision
Book a demo, ask about a pilot, or talk to the team that built it.