Fraud prevention
Automatically flag mislabelled items and correct scanning errors to ensure accurate transactions.
Industries · Retail
Vision AI for retail: product recognition at checkout and on the shelf
Streamline checkout, prevent fraud, and improve stock accuracy with automated product recognition. Tiliter’s Vision AI Agents help retailers reduce shrink, speed up service, and remove friction for customers – without relying on barcodes or manual entry.
- Faster checkout
- Fewer errors
- No barcode required
- Proven at 1,000+ stores
Powered by Object Counter, Visual Reference Agent and Label Validator
Vision AI Agents built for real-world impact
Loss prevention
Accurately distinguish similar products to enforce correct pricing, prevent losses and streamline operations
Product recognition
Speed up checkout by 90%. Instantly identify items without PLU codes or shopper input. Efficient and intuitive.
Learn more →Integration
Integrate quickly using our Developer Hub and Console. Access docs, UI/UX and tools to deploy Vision AI Agents
See it running in a live supermarket
Netto Marken-Discount uses Tiliter’s product recognition to identify non-barcoded items – fruit and vegetables – at both staffed and self-service checkouts, cutting scanning time and checkout errors.
Tiliter AI Scale
A standalone smart scale that instantly identifies fresh and unpackaged products, eliminating the need for PLU codes. Designed for frictionless customer experiences, faster checkouts and loss prevention – used in thousands of supermarkets worldwide.
Two ways to put it in the store
Run recognition on Tiliter hardware, or on the checkout hardware you already have.
As the Tiliter AI Scale
A complete unit – scale, screen, camera and printer – that drops into the produce section or a self-checkout lane.
On your existing hardware
The same recognition runs on third-party checkout hardware, and is compatible with Datalogic, Zebra and NCR.
Through the API
Send images from your own app, kiosk or trolley and get the product identity back to price it yourself.
Use cases of the Tiliter AI Scale
Scan and Go
Accelerate mobile shopping and smart trolley experiences by seamlessly identifying non-barcoded items. This reduces shopping time by 30% and helps cut down on loss and fraud.
E commerce
Increase e-commerce picking efficiency by generating barcodes readable by any handheld scanner, streamlining the order fulfilment process.
Instant price checks
Enhance the customer experience by providing instant price checks for non-barcoded items. When combined with targeted advertising, this solution boosts basket size and overall revenue.
Barcode printing
Simplify transactions and reduce loss and fraud by printing barcodes for non-barcoded items directly at the scale. The encoded barcodes can include fraud and loss prevention flags, alerting the checkout team when necessary.
Trusted by industry leaders
Used in thousands of supermarkets worldwide, including more than 1,000 stores across Australia and New Zealand.
1,000+Stores deployed
50M+Images processed every year
90%Faster checkout on non-barcoded items
Retail questions, answered
How does produce recognition work at a self-checkout?
A camera at the checkout or on the scale looks at the produce the customer has actually put down and identifies the variety, not just that it is fruit, so it is rung up without anyone hunting for a code. It takes about 300 milliseconds, which is the part that matters when someone is standing in the lane.
Can it check the shelf, or only the checkout?
Both. Shelf and stock checks run as scheduled workflows: someone photographs the aisle or the bay on their phone, and the result comes back as a count against what should be there, recorded with the image. That is the object counting agent doing the work rather than recognition at the till, which is why the two sit on the same platform.
Do customers still have to look up PLU codes?
That is the step this removes. PLU lookup is where loose produce goes wrong: the customer picks the wrong variety, or picks the cheapest thing that looks close, and the store carries the difference. Recognising the item on sight takes the decision away from a chart on a screen.
What is the Tiliter AI Scale?
A scale with produce recognition built in, so fruit, vegetables and other weighed items are identified as they are weighed rather than picked from a list. It is one of the two ways to put this in a store. The other is recognition running against cameras on the checkouts you already have.
Can it fetch product images into our POS automatically?
Yes. Identification returns the product identity along with structured data, which is what a POS needs to attach the right image and description to an item instead of someone maintaining that mapping by hand. See product recognition for how the matching works and where the output can be sent.
How does it reduce shrink and misscans?
Most loss at self-checkout is not theft. It is the wrong item being entered: a premium variety rung up as a cheap one, deliberately or not. Identifying the item visually removes the opportunity, and every check leaves a record with the photo attached, so patterns become visible instead of guessed at.
Do we have to replace our checkout hardware?
Not usually. Recognition can run against lanes you have already fitted out, including Zebra, Honeywell and Datalogic units, or you can put in AI Scales where weighed items are where the money leaks. What is already installed tends to decide it.
Does it work for packaged goods as well as fresh items?
Yes. Fresh is the harder case, because a loose apple looks different every time, but the same recognition covers packaged products, and identifying parts and assets outside retail works the same way. Fresh is simply where it had to be proven first.
Which retailers are using it?
It runs across more than 1,000 locations and processes over 50 million images a year, with grocers in Australia and Europe among them. The logos on the home page name the ones that are public.
Ready to turn images into operational data?
Build visual verification workflows and deploy them through mobile, web, cameras or your existing systems. Deployed in thousands of locations worldwide and ready to scale with your operations.