Automating manual processes with AI capabilities for Digital GoToMarket

  • Customer
  • Industry
  • Service
  • Segment
  • Author
  • Digital GoToMarket
  • Information Technology
  • AI/ML
  • SMB
  • Jodie Rhodes
Explore more
At a glance

The Digital GoToMarket team helps companies craft digital strategies and launch successful subscription products and services.

Challenge

Digital GoToMarket (DGTM) faced challenges automating data extraction, as rigid expressions limited nuanced discovery.

Solution

Firemind built a POC using AWS to create a scalable, ML-driven document automation pipeline.

Services used
  • Amazon Comprehend
  • Amazon Textract
  • AWS Lambda
  • Amazon S3
Outcomes
  • Automated IDP, reduced manual effort.
  • Expanded capabilities, freed up human operators.
Business challenges

Limitations of manual data extraction processes

Currently, the DGTM collection, and extraction of data, from the varied digitised documents, is all gathered with ‘regular expression’ – a very explicit method of finding and extracting data from the documents.

This had proven to be reliable to a point, however it did not allow for discovery of more nuanced information, and it also required a very specific set of requests be generated for documents, in each of the backed securities markets.

DGTM wanted the capability to apply a machine learning strategy to this document analysis workflow. Providing multiple benefits that included scalability and adaptability, as well as an evolving level of accuracy that can be leveraged to improve data models over time.

What our customers say

Hear directly from those who’ve experienced our services. Discover how we’ve made a difference for our clients.

Solution

Leveraging AWS for automated document processing

Our solution was to build out a proof-of-concept (POC) to demonstrate the capability of AWS Textract and AWS Comprehend for a document handling and analysis pipeline. The 2 focused goals of this solution were to provide an ML ‘sandbox’ to validate and test analysis model (custom classifications) as well as an automated pipeline for document handling and model retraining.

The ‘sandbox’ is a set of resources designed to provide an automated pipeline of AWS resources and services. This allows for a batch of documents to be uploaded, processed and analysed, so that the output can be reviewed and results used to influence the training of the custom classifications.

Firemind were able to remain true to the projected timeframe of 22 days, set over 7 phases. The proof-of-concept worked to provide accurate automation of previous manual tasking, enabling an environment to start moving real data into production.

Automated processes

The POC provided an automated process for Intelligent Document processing (IDP). Saving cost in manual hosting and storage as well as time, freeing up human operators to work on other tasks.

Introduction to managed services

DGTM were now using machine learning services such as Amazon Comprehend and Amazon Textract for improved extraction and classification during the processing logic.

Reduced human input

Removing the need for staff to work on these processes enabled the expansion for edge cases and faster classifications across the entire process. Human operators were now freed to work on other areas of the business and support their customers in other areas.

Get in touch

Want to learn more?

Seen a specific case study or insight and want to learn more? Or thinking about your next project? Drop us a message!