Image-Based Product Identification and Catalogue Candidate System

The problem

In day-to-day operations, customer photos and product catalogue data arrive in separate records and files, preventing the team from reviewing the same case in a single view.

As a result, the link to source data may be lost while preparing a list of possible products and variants, leading to repeated checks and uncertainty about the basis for the decision.

Incomplete automation may overlook poor image quality or the presence of several strong candidates, while unsuitable automation carries the risk of ordering the wrong variant.

What we set out to improve

• Time taken for the draft list of possible products and variants to be ready for review
• Correct routing of records to the appropriate specialist queue when image quality is poor or there are several strong candidates
• Logging of suggestions amended or rejected by the catalogue specialist

How the system works

1. 01 Retrieve the customer photo and product catalogue data from an authorised source
2. 02 Verify the source system, record identifier and recency details
3. 03 Transform the fields into the target process schema
4. 04 Prepare a draft list of possible products and variants
5. 05 Compare the draft with business rules and source records
6. 06 Check for the exception: image quality is poor or there are several strong candidates
7. 07 The catalogue specialist approves, amends or rejects the draft
8. 08 Write the input, suggestion, amendment and final decision to the audit log

Methods we used

• Source-traceable data extraction for the customer photo and product catalogue
• Structured output that constrains the fields in the list of possible products and variants through business rules
• An exception gate for poor image quality or the presence of several strong candidates
• A queue showing the source and suggestion together for the catalogue specialist
• Duplicate-processing prevention and rollback records to manage the risk of ordering the wrong variant

Where people stay involved

Data and security

Only the fields required for the task are processed from the customer photo and product catalogue. Source-system permissions are preserved, and personal and commercially sensitive data is minimised before being passed to the model. Every read, suggestion, human amendment and write to the target system is logged.

Who this suits

Frequently asked questions