Retail
with Vision

Better customer experience. Less checkout loss. Powered by item-level recognition at scale.

better shopping experience

A $2 mistake  
shouldn't stop
the lane. We know it was a box of crackers, not a steak.

The shopper gets a nudge, not an audit. One tap and it is corrected. No flashing light, no queue behind you, nobody walking over, no awkward confrontation.

Book a demo

How we improve and secure the checkout experience

Self-checkout lane camera feed
Live
08:24:39
CAM 04
4k
60FPS
80 MBPS
Detected

Salmon Fillet

$12.99

Salmon fillet
SKU 184-027
Scanned

No matching item

Self-correction
Self-checkout screen showing correction of a mis-scanned item.

Mis-Scan

Item passes over the scanner with no matching scan.

Self-checkout lane camera feed
Live
08:24:39
CAM 04
4k
60FPS
80 MBPS
Detected

Premium Steak

$48.99

Premium steak
SKU 184-027
Scanned
Red strawberries

Red strawberries

$4.99

SKU-42371

Ticket Switch

Premium steak scanned as strawberry. $19 differential.

Self-checkout lane camera feed
Live
08:24:39
CAM 02
4k
60FPS
80 MBPS
Detected

Pineapple

$4.47

Pineapple
SKU 184-027
Pos record
Onion

Onion

$2.99

Correct the mismatch
Self-checkout screen correcting a produce-switch mismatch.

Produce Switch

Pineapple weighed, onion entered. Cheaper item rung.

Self-checkout lane camera feed
Live
08:24:39
CAM 02
4k
60FPS
80 MBPS
Detected

Northstar cola 0,5L

$2.99

Northstar cola bottle
SKU 184-027
Correct the mismatch
Self-checkout screen correcting an item left in the shopper’s hand.

Left in Hand

Unscanned item past scanner. Caught at bagging.

Loss capture
70%
SKUs · item-level
100K+
Millisecond response
<100
False alerts
<5%
ITEM-LEVEL RECOGNITION

We know the
 item.
Gesture systems guess.

Weight and motion tell you something moved. Only item-level recognition tells you what it was and what it is worth.

how it works
Gesture system
Overhead self-checkout camera view used by a gesture-based system.
Live
08:24:39
CAM 04
motion event
TRANSFER DETECTED
SCAN ZONE → BAGGING ZONE
ZONE 02      −0.46 KG
ZONE 03      +0.46 KG
DURATION     0.38 S
ITEM ID      UNKNOWN
EVENT-6812
ShopicShopic
Self-checkout lane with Shopic identifying each individual item.
Live
08:24:39
CAM 04
strawberries
plus icon
42371
eggs
plus icon
42371
spinach
plus icon
42371
pos match
Briar Meadow strawberries in a clear clamshell punnet.
BRIAR MEADOW
$4.79
PACKAGING CLEAR CLAMSHELL
SKU SKU-42371
Gesture-based
Shopic
Method
Motion only
Item recognition
What the shopper feels
Frequent false stops
Minimal Interruption
Intervention
Untiered intervention
Optimal intervention
What the store gets
Narrower coverage
Full assortment
Processing
Requires server
Edge-first
A feature only we offer

Optimal
Intervention. The right action at the right time.

Because we know the item and its value in every event, the response tunes itself to what's actually at risk.

Low Risk
High Risk

Self-correction

A gentle prompt. The shopper fixes it themselves.

Shopper at a self-checkout tapping the screen to add a missed item
unscanned item
Broccoli
value
$1.28
nudge

Attendant assist

Floor staff gets a targeted, evidence-backed alert.

Store attendant checking an alert on a self-checkout screen
item staged to bagging area
Bottle of laundry detergent
detergent 2l
$18.99
Apple
not: apple
attendant assist

Transaction hold

High-risk events are stopped before they complete.

Store attendant stepping in at a self-checkout while a shopper holds a banana
ticket switch
Packaged ribeye steak
ribeye
$48.99
Banana
not: banana
transaction hold
Web app

What happens at the store - all in our ShopWatch app

Every incident, every pattern, at every store.

Compare shrink rate across every location, down to the day and the hour it happens.

See exactly which items are targeted, and how often.

Capture rate climbs and false positives drop as the system learns your store.

See the portal
ShopWatch dashboard showing estimated loss potential, active learning coverage, and recent shrink attempts
WHAT THE STORE GETS

Because the shopper isn't interrupted,
the store isn't either
.

A shopper leaving the store with her groceries, waving to a member of staff
Live
08:24:39
CAM 04
Smiley face
smooth checkout
time saved
(vs. typical checkout)
00:12
staff interventions
0
happy customers
1
the foundation

Five years of item-level data that
nobody else has
.

100,000+ SKUs. Seven transactions to learn a new item. Millions of real checkout events across three continents.

Shopic edge camera used for visual product recognition.
01 — YEARS OF ALGORITHM RESEARCH AND DEVELOPMENT

Item-level recognition at grocery scale is not available off the shelf. No other company in the market managed to deliver similar results.

5+
02 — ANALYZED TRANSACTIONS

Real shoppers in real stores, training data that cannot be bought or simulated.

2,986,987
03 — VALIDATION AND INTEGRATIONS

Years of validation in live stores, plus 8+ years of POS integrations, NCR Voyix, GK, Diebold-Nixdorf, and proprietary stacks.

Live‑store
04 — PATENTS GRANTED WORLDWIDE

Granted in the US, Israel, Canada, Europe, and China — covering the device, the models, and the method.

5+
01 — YEARS OF ALGORITHM RESEARCH AND DEVELOPMENT
5+
02 — ANALYZED TRANSACTIONS
2,986,987
03 — VALIDATION AND
INTEGRATIONS
Live‑store
04 — PATENTS GRANTED WORLDWIDE
5+
one compounding deployment

ShopWatch
 is just the start. You're putting
a platform
 into every store.

Every new application reuses the same catalogue, the same edge box, the same integration. Each one shortens the time-to-value of the next.

see the platform
Isometric grocery store with aisles, shoppers, self-checkouts, and a produce section.Shopic AI engine connecting store activity with item-level data.Item-level product data foundation represented by a grocery catalog.
03 APPS

Store Apps

Instant Checkout
ShopWatch
Smart Cart
Picklist Assist
Shelf Monitoring
02 ENGINE

AI Engine

Shopic IQ
Vision AI Models
Scene Understanding
Continuous Learning
01 FOUNDATION

Item-level Foundation

Item Data
Visual Catalog
Store Interactions
Event Data
FAQ

Straight answers, for people and for machines.

What is Shopic?

Shopic builds item-level computer vision for grocery. One camera and one edge box turn a normal checkout lane into something that knows what every item actually is.

ShopWatch is the loss-prevention application that runs on that foundation. It watches self-checkout lanes item by item and flags only the events that are worth a person's time.

Recognizing the specific product, not just that something moved. Weight and motion tell you a transfer happened; item-level recognition tells you it was a ribeye and what it is worth.

Yes. Shopic integrates with the major POS stacks — NCR Voyix, GK, Diebold Nixdorf and proprietary systems — on top of eight years of integration work. No lane replacement.

Both, plus produce switching and items left in hand. Because the system identifies the product itself, a swapped barcode simply does not match what the camera sees.

Smart cart is one application of the platform, not the company. The same catalogue and edge processing also power checkout, picklist assist and shelf monitoring.

A shopper selecting fresh produce in a grocery store.
Live
08:24:39
CAM 04
+
Staff interventions today
0
Auto recognized
Honey crunch
$3.49
Sunspire Honey Crunch cereal box.
CONTACT US

Want to understand how the store operates?

See your own loss baseline before you commit. Shadow mode runs two to four weeks and measures your real numbers.