Digital GoToMarket (DGTM) helps companies craft digital strategies and launch successful subscription products and services. They work across industries to optimise operational efficiency and deliver measurable business outcomes.
Challenge
DGTM’s document data extraction process relied on rigid regular expressions, which worked for structured patterns but failed to capture nuanced information. This method also required complex, document-specific configurations for different markets, limiting scalability and adaptability.
They needed a more intelligent and flexible approach – one that could evolve over time, improve accuracy, and streamline processing. By introducing machine learning into their workflow, DGTM aimed to reduce manual effort, speed up analysis, and improve the overall quality of extracted data.
Solution
Firemind developed a proof-of-concept using AWS to replace manual methods with a scalable, ML-driven document automation pipeline. The system used Amazon Textract for advanced text extraction and Amazon Comprehend for intelligent classification, enabling batch uploads, automated analysis, and iterative accuracy improvements.
An “ML sandbox” environment was created for testing and retraining models, allowing DGTM to adapt their extraction process over time. Deployed in just 22 days, the Intelligent Document Processing (IDP) solution freed up staff for higher-value tasks and provided a strong foundation for scaling automation in the future.
Services used
- Amazon Comprehend
- Amazon Textract
- AWS Lambda
- Amazon S3
The Results
- Automated document processing, reducing manual input
- Increased accuracy and adaptability of data extraction
- Frees up human operators for higher-value tasks
- Provides a scalable foundation for future ML capabilities
- Delivered in just 22 days
