success story

Streamlining food label verification with AWS GenAI-powered automation

A family-owned American chain of convenience stores faced significant challenges with its food label verification process. The manual approach was labor-intensive, time-consuming, and prone to errors. With over 200 items needing verification daily at each location, the client saw an opportunity to improve efficiency and accuracy through AWS genAI and automation. Virtusa implemented a cutting-edge solution to streamline the verification process, ensuring product information matched the source data accurately.

Ensure that you open the links to enable the Next button, assuming you have read all the documents.
Okay & NextClick to Next
The Challenge

The client's existing process required manual verification of key attributes on food labels, such as Product Description, Net Weight, Allergens, and Lot Code. This method was slow, error-prone, and inefficiently used valuable labor resources. To enhance productivity and reduce potential mistakes, the client sought a more reliable, automated system that could handle large-scale daily verification tasks.

The Solution

Virtusa developed an automated label verification tool leveraging AWS genAI services, including Amazon Bedrock and Amazon Textract. This genAI-powered solution automatically extracted crucial information from product label images, such as Product Description, Bioengineered Statement, Enjoy Through Date, and Lot Code—and compared it against a predetermined truth dataset to ensure consistency and accuracy.

The tool generated comprehensive reports detailing the matching rates and flagged discrepancies between the extracted and reference datasets. Additionally, it meticulously tracked performance metrics such as token usage, processing time, and an overview of image processing steps. These insights allowed for ongoing refinement and optimization of the verification process.

Virtusa implemented genAI models to boost the system's effectiveness as part of the solution. The client learned that the quality of results depended on the genAI models used—more advanced and latest models resulted in higher accuracy rates and better overall performance. With continuous improvements to the system, the tool achieved a matching rate of over 95% for non-complex labels, significantly reducing the error rate and manual intervention required for failed matches.

AWS services used:

  • Amazon Bedrock
  • Amazon Textract
  • AWS Lambda
  • Amazon API Gateway
  • Amazon DynamoDB
  • AWS Storage Gateway
  • Amazon S3
  • Amazon EventBridge
  • Amazon SQS
  • Amazon CloudFront
  • Amazon Route 53
  • Amazon VPC
Streamlining food label verification with AWS -Solution
The Benefit

The implementation of Virtusa's AWS genAI-driven solution provided several key benefits for the client:

  • Time efficiency: The automated tool dramatically reduced the time required for label verification, enabling the client to quickly process over 200 items daily per store. 
  • Labor optimization: By automating the verification process, the client could reallocate labor resources to more productive, business-critical activities. 
  • Enhanced accuracy: Achieved a ~95% accuracy rate for non-complex labels, ensuring reliable verification with fewer manual reviews.
  • Improved process insights: The tool's tracking of key performance metrics provided valuable insights, further optimizing the verification process and improving operational efficiency.
Reimagine your business with enterprise-wide cloud adoption

Realize the true value of AWS

Learn more Click to Learn more
Related content