Dynamo Software Inc. Enhances Document Classification with AI and Automation

Dynamo Software Inc., formerly known as Netage Solutions, is one of the world's leading cloud providers of alternative investment management software. Dynamo specializes in premium, industry-specific, configurable asset management and reporting software for the alternative assets industry. Its products and services cover markets such as private equity and venture capital funds, real estate investment firms, hedge funds, funds of funds, prime brokers, foundations, endowments, pension funds, and family offices. The Dynamo platform is an intuitive and highly configurable, end-to-end cloud solution that improves the productivity of fundraising, deal, research, investor relationship, and portfolio management teams worldwide.

Dynamo Software Inc., a leading cloud provider of alternative investment management software, was seeking to enhance its document classification platform through AI and automation. The platform was designed to store, classify, and transfer information and metadata from various documents to appropriate investments. However, Dynamo wanted to improve the accuracy of document classification and gain the ability to make predictions based on a document's content. The goal was to reduce the amount of repetitive manual work performed by their data team, lower operational costs, increase performance, and minimize the time needed for making decisions on client investment portfolios. The existing platform received thousands of various types of documents every month, some of which were manually added by managers. Dynamo wanted to significantly improve the accuracy of their existing ML tool, automate a portion of the data processing pipeline, and achieve at least 85% accuracy on new data.

Provectus, an AWS Premier Consulting Partner with competencies in Machine Learning and Data & Analytics, was selected to join the Dynamo project. Provectus set up an infrastructure for Dynamo's development and management environments, as well as an experimentation infrastructure for document classification. They conducted an exploratory data analysis (EDA) on Dynamo’s datasets to develop a testing dataset and data extraction datasets, and built a baseline classification model. A robust pipeline for document classification was implemented. The document classification model designed and built by Provectus exceeded Dynamo's expectations, returning an f1-score of 95% on a test dataset. The training pipeline was built using AWS best practices for data security and privacy, ensuring data confidentiality. The suite of Amazon SageMaker services was used to develop the model building pipeline, while the inference pipeline was built on AWS Step Functions.

• The new document classification solution developed by Provectus for Dynamo demonstrated the potential of AI and automation. The solution was able to automatically classify over 90% of various documents with AI and make predictions based on extracted document data. This success incentivized the leaders of Dynamo to move forward into the next stage of cooperation with Provectus. The solution not only reduced the amount of repetitive manual work performed by their data team but also lowered operational costs, increased performance and productivity, and minimized the time needed for decision-making on client investment portfolios. Dynamo is now ready to move forward into the next stage of cooperation with Provectus and AWS.

• New document classification solution delivered in six weeks

• 95% accuracy on a customer test dataset of PDF documents

• Exceeded the requested accuracy threshold of 85%