Case Studies | Retail

Reading the label, automatically

LLM Product Label Validation PoC

About

A Retail business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The challenge had three focus areas, each one a step from a raw image towards a compliance decision.

Reading the label reliably

Product labels are dense, curved and visually busy, mixing headings, tables, small print and symbols. The first focus was to read the text off a photograph of the packaging accurately enough to trust what came out.

Structuring the extracted content

Raw text is not enough. The extracted words had to be organised into the fields that matter for compliance, the nutritional information table, the

Validating against market rules

The end goal was compliance. The final focus was to show that once a label had been read and structured, generative AI could reason over that data and flag where it did or did not meet the labelling requirements of another country.

Solution

A photograph of the product label is passed through optical character recognition, which reads every line of text on the packaging and returns it with a confidence score and its position on the label. On the sample labels this recognised the core fields, from the nutritional information table and reference values through to the full ingredient list and pack size, with high confidence.

99%+

Text recognition confidence on core label fields

3

Supplement lines tested: turmeric, ashwagandha, multivitamin

PoC

Image to structured, checkable label data

By the numbers:

  • 99%+ - Text recognition confidence on core label fields
  • 3 - Supplement lines tested: turmeric, ashwagandha, multivitamin
  • PoC - Image to structured, checkable label data
Changes

The proof of concept met its goal. Working from real photographs of the client's products, the system read the labels, structured the content and demonstrated that generative AI could reason over the result for compliance, exactly the end-to-end path the engagement set out to validate. The approach was proven across several of the client's supplement lines.

  • Accurate reading provenThe nutritional information, reference values, ingredients and pack details were lifted from label photographs with high recognition confidence.
  • Structured output deliveredExtracted text was organised into the fields that matter for a compliance check rather than left as raw words.
  • Compliance reasoning demonstratedGenerative AI on Amazon Bedrock showed it could reason over the structured label against another market's requirements.
  • HandoverThe client left the proof of concept with clear evidence that automated, image-to-compliance label validation is viable for its range.

With viability established, the natural next steps (projected) would be to widen coverage across the full product range, encode the labelling rules for each target market, and wrap the flow in a simple interface so compliance checks can be run routinely as new products and markets are added.

AWS Stack

Amazon Textract

For optical character recognition that reads the text off product label images.

Amazon Bedrock

For the large language model that structures the extracted text and reasons over it.

Amazon S3

For storing the source label images and the extracted output.

AWS Lambda

For the serverless processing that ties the extraction and validation steps together.

YOU MIGHT LIKE

Related success stories

View all case studies

Case Studies | Insights

Utilising Language Recognition, Speed, and Enhanced Security to Make Social Media a Force for Good

  • Here, we take a detailed look at how the Cloud Combinator team collaborated with another cutting-edge AI service provider that provides intelligent systems to “make social media more social” for brands and users alike.
  • Arwen AI is a UK-based startup specialising in AI solutions to manage and enhance brands’ social media interactions. Founded in 2020 by Matt McGrory, Dr. David Cole, and Joel Bailey, Arwen. AI focuses on using AI to automatically detect and remove spam, toxic comments, and other unwanted content from social media platforms.
  • The team at Arwen have three core products. ‘Moderate’ is focused on identifying and removing toxic content from social media channels. ‘Engage’ helps brands identify and engage with meaningful conversations on social media, and ‘Customize’ allows brands to apply bespoke algorithms to their channels - creating an even more effective moderation and engagement.
Read more
CONTACT US

Ready to turn AI into impact?

We'll help you spot the highest-value opportunities, reduce risk around your first AI initiative, and define a clear path to results from day one.

Why talk to us:

Outcome-driven recommendations

AWS-recognised delivery expertise

Risk-aware AI adoption

Clear next step, not a sales pitch

Start with a focused 20-minute conversation about your goals — no pressure, no commitment.

This website uses cookies to enhance user experience and to analyze performance and traffic on our website.

See our Privacy Policy for details.