Skip to main content

Focal

Focal Systems · Whitepaper

The ROI of
Shelf AI

How AI-powered shelf cameras are delivering measurable returns for grocery retailers.

99%
On-shelf availability
Proven improvement
2–5%
Availability increase
Across deployments
200M+
Daily product scans
Across the platform
23M+
Gaps filled
And counting

Executive Summary


Grocery retailers operate on razor-thin margins — often just 1–3% net profit — yet they lose billions annually to a problem hiding in plain sight: empty shelves.

Shelf AI, the application of computer vision and deep learning to continuous, automated shelf monitoring, is changing the economics of on-shelf availability. By deploying AI-powered cameras that scan every product on every shelf every hour, retailers gain real-time intelligence that transforms replenishment, ordering, inventory accuracy, and labor productivity simultaneously. This paper quantifies that value through real-world deployments at Morrisons and Greenfield’s ShopRite.

01The Cost of Empty Shelves


Inventory distortion — the combined impact of out-of-stocks and overstocks — costs the global retail industry $1.7 trillion a year, about two-thirds of it from out-of-stocks (IHL Group, 2026). And the shopper impact is just as steep: more than 30% of shoppers who hit an empty shelf leave for a competitor.

$1.7T
Inventory distortion
5% of retail revenue globally
67%
Driven by out-of-stocks
Share of that loss
30%+
Shoppers leave
For a competitor
$7.4B
Out-of-stock churn
Grocery Doppio, 2025

02How Shelf AI Works


Shelf AI uses AI-powered cameras mounted directly on store shelves to continuously monitor product availability, detect out-of-stocks, identify low stock levels, measure planogram compliance, and feed real-time data into replenishment and ordering systems.

Unlike periodic manual scans or robot-based solutions, shelf-mounted cameras provide continuous visibility — scanning every product on every shelf every hour of every day. Focal Systems’ implementation consists of four connected components:

STEP 01

Computer Vision

Shelf cameras scan hourly at >95% accuracy, detecting outs, lows & non-compliance.

STEP 02

Shelf AI Engine

40M+ scans/day; prioritizes the gaps that matter most by value.

STEP 03

Action Tool

Prioritized task lists sent to associates’ handheld devices.

STEP 04

Impact Dashboard

Measures labor, availability, compliance & sales recouped.

03Case Story: Morrisons


“This system is bringing down the in-day replenishment times significantly, which in turn is having a positive effect on availability, sales and customer satisfaction.”

Rami Baitiéh, CEO, Morrisons

Morrisons, one of the UK’s largest supermarket chains operating approximately 500 stores, deployed Focal’s Shelf AI in a six-month rollout — one of the largest computer-vision deployments in retail globally. With 400–600 cameras per store scanning shelves hourly, the CEO described it as a game changer for both productivity and availability, with the retailer now among the UK’s top performers for on-shelf availability.

The partnership won The Grocer Gold Award 2025 for Technology Initiative of the Year. Morrisons trialed multiple providers, and the results showed the best solution was with Focal — in the CEO’s words, store teams asked them “please don’t go back to the old systems.”

Hear from Morrisons

“The implementation of Focal Systems has fundamentally changed how we operate our replenishment operation and how we manage our stock records.”

Gordon Macpherson, Productivity Director, Morrisons

04Case Story: Greenfield’s ShopRite


“We are in the customer happiness business; we want to eliminate all points of disappointment. With Focal, OSA is almost 99%.”

Seth Greenfield, Owner, ShopRite

Greenfield’s ShopRite deployed Focal’s Shelf AI to achieve near-perfect on-shelf availability — and reached almost 99%. For context, industry averages run 92–95%, so a 4–7 point improvement is a substantial competitive advantage.

92–95%
Industry avg
~99%
Greenfield’s ShopRite

Extra Added ROI

Focal’s Theft Spotter flags stock that vanishes without a matching sale — pinpointing the hour, location, and product taken — so the loss-prevention team at Greenfield’s ShopRite catches repeat offenders roughly 70% of the time, versus sifting through hours of footage before.

“Annual shrink is in the millions of dollars a year; on the theft-alerting side, it’s thousands of dollars a week that we’re catching.”

Seth Greenfield, Owner, ShopRite

05The Operational KPIs That Drive ROI


Availability

Continuous scans detect out-of-stocks in near real-time; prioritized replenishment restocks the highest-value items first.

Impact: 2–5% availability increase; almost 99% OSA at Greenfield’s ShopRite.

Productivity

Eliminates hundreds of hours per store per month of manual scanning; associates get automated task lists.

Impact: Zero manual scanning hours; associates redeployed to customers.

Inventory Accuracy

Continuous scans update balance-on-hand in real time, correcting phantom inventory.

Impact: Accuracy raised to >95%; 100,000+ adjustments daily.

Ordering & Planning

Accurate, real-time shelf data feeds computer-generated ordering, ending faulty over- and under-ordering.

Impact: Excess inventory reduced by 30%+ while improving availability.

Shrink Reduction

Shrink comes mainly from theft and spoilage. Theft Spotter flags stock that vanishes without a sale, while prioritized replenishment moves perishables onto the floor before they spoil.

Impact: 2x+ theft identification; less spoilage.

The question is no longer whether Shelf AI works. It’s how long retailers can afford to operate without it.

With 300,000+ cameras deployed globally and 40 million shelf scans processed daily, the barriers to adoption have never been lower.