Hyperscience: Delivering the much-needed AI and ML-enabled document processing solution for government entities

Unlike the private sector entities who are adept at leveraging newer technologies, most government agencies still rely on legacy systems and repetitive, manual data entry, which is slow, expensive, and error-prone, resulting in poor outcomes for agencies as well as citizens. Existing document processing, in particular, typically involves a lot of manual work and clunky legacy software that fails to reliably read and transcribe realworld, messy documents—handwritten applications or claims, invoices, or distorted scans or faxes. Most government agencies are stretched like never before, faced with increased workloads and data backlogs that need to be processed accurately and efficiently with constrained budgets and limited staff resources.

This data and processing bottleneck makes it difficult to respond to citizens or provide critical services in a timely fashion. Government agencies need to find ways to embrace AI and ML and other cutting-edge technologies to keep up with demand and service citizens. That’s where Hyperscience can help.

Hyperscience automates most of what they call the critical “step zero” of the business process —data entry—with its Intelligent Document Processing (IDP) solution. “We use cutting-edge machine learning techniques to automate document processing and data extraction with higher accuracy than previously possible, and the solution continues to learn on an organization’s data—behind their firewall—to drive lower error rates and greater automation,” mentions Peter Brodsky, CEO and co-founder of Hyperscience.

Since 2014, Hyperscience has focused on perfecting the mission-critical component of its platform, using cutting-edge, proprietary AI and machine learning techniques that have been developed and trained for over five years. The solution automatically classifies, separates, and deduplicates diverse documents, and then extracts and exports data from those documents for downstream processing and analysis. This is particularly relevant with NARA mandates to store and transfer all records in an electronic format over the next few years. Government agencies have had to look at document processing solutions to remain compliant, and no solution is as accurate as Hyperscience when it comes to reading messy, diverse documents and extracting and structuring the relevant data agencies need so that it can be further analyzed.

Hyperscience is built to read messy, real world documents and extracts and structures the relevant data agencies need so that it can be digitized and analyzed

Hyperscience works with many of the world’s leading government, financial services, insurance and healthcare organizations, including TD Ameritrade, QBE, Mutual of Omaha and Voya Financial, helping them to lower costs, reduce error rates by 67 percent, increase employee capacity by 10x and improve customer experience. For example, with Hyperscience, TD Ameritrade was able to double processing speed with 50% fewer data keyers, quadrupling throughput within the first four weeks. The client ended up completing a one-time compliance project, part of their acquisition of Scottrade, within four months – one-third of the time they originally budgeted. They quickly turned their attention to using Hyperscience to support other content digitization efforts, and today they’re using the technology to process financial documents and extract data with 98% automation and 99% target accuracy.

Among leading automation solution providers, Hyperscience is the only one with a proprietary machine learning core. This allows the solution to read and extract data from documents (with all their imperfections) that would cause other technology or legacy solutions to fail. “Our built-in quality assurance mechanism helps the machine establish truth in the accuracy of its extraction,” adds Brodsky. “This not only provides visibility into the machine and human performance but also helps fine-tune the underlying models, enabling customers to unlock throughput improvements on Day 1 and continue to get better over time.” What’s more, it can be deployed on-premises or within the client’s private cloud instances, so data never leaves their environment, ensuring complete control over sensitive information.

Hyperscience has witnessed tremendous traction over the last five years, and in June 2020, secured a $60 million Series C led by Bessemer Venture Partners with major investment from Tiger Global Management. The funds are being used to fuel product development and market expansion, including within the federal government space and continued international expansion into the UK & EMEA. The company continues to grow its federal sales operation and is seeing increased demand for automation. Usage of the Hyperscience Platform has increased at a rate that is 3x greater than what it was before the pandemic, as enterprises everywhere are looking to become more efficient amongst changing market dynamics and economic uncertainty.