Skip to content

Product recognition by photo

Customers show their product instead of searching for model names and numbers.

One photo is enough: the AI finds the matching product in your catalog and reads the product, article and serial number from the rating plate. It works by phone, by email, in chat or in your own app. You add new products without any model training.

Which model do you have exactly?
The hardest question in customer service.

Customers rarely know the exact model name of their device. Rating plates are small, dirty or hard to reach. Long numbers get misspelled on the phone or mistyped in forms. This leads to follow-up questions, longer handling times and, in the worst case, wrong spare parts being shipped. With a photo, it is clear which product the customer is holding.

How it works

1. Send a photo

The customer takes a photo of the device or the rating plate, via a link on their smartphone, as an attachment or directly in your app.

2. Recognize the product

The AI compares the photo with reference images from your catalog and reads the numbers from the rating plate at the same time.

3. Use the result

You receive the most likely products together with product, article and serial number. Your system or your team continues working with them right away.

Your benefits

New products without training

A new product can be recognized right away with just a few reference photos. Image recognition models don't need to be trained at all.

Image and rating plate cross-checked

Image recognition and rating plate are evaluated at the same time. If they contradict each other, the system asks instead of guessing. Devices without a reference image are also recognized via the rating plate, and QR codes and barcodes confirm the numbers.

Photos are not stored

The photo is only processed for recognition and discarded afterwards. Only the result ends up in the conversation.

Your use cases

Voice agent

In the middle of the call, the agent sends a link by text message. The caller takes a photo while the conversation continues. You will find an example further down.

Email and ticketing system

Photos from emails or contact forms are analyzed automatically. The ticket immediately shows the recognized product and the numbers.

Integrates with your applications

Our image recognition APIs can also be integrated into other apps and use cases, such as your web shop, chat or field service app. All they need is the image.

Use case:
Spare parts search in the voice agent

A customer calls because they need a spare part but don't know their model. This is how product recognition works during the call:

1. Recognize the request

The agent recognizes that it needs to know the exact product, for example for a spare parts order, a complaint or a warranty request.

2. Link by text message

With the caller's consent, the agent sends a text message with a link to a lightweight web page. The caller doesn't need to install an app.

3. Take a photo

The caller photographs the device or the rating plate directly in the browser. The call stays open.

4. Recognize the product

The AI compares the photo with reference images from the product range and reads the numbers from the rating plate at the same time.

5. Confirm the selection

The caller sees the most likely products with images and taps the right one. If it isn't there, one tap on “None of these” is enough.

6. Back to the conversation

The result goes straight back into the ongoing call. The agent suggests matching spare parts and takes the order.

What's behind it:
proven building blocks

The image recognition components run in Microsoft Azure and are connected via APIs.

Voice agent

The conversation is handled by a voice agent on NiCE Cognigy. It opens the photo page directly from the call.

Image recognition without training

AI image features compare the photo with your product catalog. When your range grows, you only add reference photos.

Reading the rating plate

Text recognition reads the numbers from the rating plate and also evaluates QR codes and barcodes.

Experience the live demo

We show you product recognition using the example of a spare parts search by phone and discuss how it can be integrated into your customer service and your applications.