Intelligent Grants Automation: Unlocking the Potential of Machine Learning
What was once only the stuff of sci-fi movies has now become a cornerstone technology for government agencies to help reduce inefficiencies, improve services, and be dramatically more effective in responding to citizen needs. At Dovel, we have been working closely with our customers to identify how machine learning can be applied to solve some of the most complex problems facing their businesses.
Machine learning (ML) is an application of artificial intelligence that uses computer systems to “learn” about specific tasks without being explicitly programmed to do so. ML is found in the products we use every day, from a bank alerting us to potential fraud, an email provider filtering out spam from unwanted sources, or a streaming video service recommending which movies we might like to see. In all of these scenarios, systems are identifying patterns and making predictions based upon data, whether it be our penchant to frequently shop at the same stores or our preference for action movies over romantic comedies.
These same capabilities have the potential to make transformational improvements to the business operations across all government agencies at all levels. ML can help automate monotonous tasks and make predictions that feed a more efficient and effective government. The opportunities to do so are endless, from improving customer service through chatbots to predicting when components may fail on combat equipment to identifying potential identify fraud at our borders.
Dovel has been working closely with HHS GrantSolutions to utilize machine learning to help improve the return on investment of the more than $700 billion in Federal grant programs. GrantSolutions is an award-winning, government-designated Software as a Service (SaaS) provider that manages more than $80 billion in annual Federal grant awards for grant programs across HHS, DHS, DOT, EPA, HUD, Treasury, SSA, and the Denali Commission. In 2017, Dovel helped GrantSolutions launch a pilot program to explore how machine learning bring together data from multiple sources to help agencies decide where to best direct grant investments and to assist them in monitoring and overseeing grant projects.
The processes that govern grant investments require an immense administrative burden on both grantees and Federal agencies. Multiple surveys and studies have shown that grantee organizations spend more than 40% of their grant resources on administration activities alone. Federal agencies report similar frustrations combined with a reduced workforce available to oversee and execute the grant programs. This means that substantial resources are being diverted from critical program mission objectives.
The initial applications of machine learning in GrantSolutions have resulted in new tools that are dramatically reducing the burden for agency grant managers and giving them better insight into the organizations they are funding. Using data from government systems and other public sources, Dovel is identifying patterns that help predict potential behaviors and outcomes. For example, the tools can predict the likelihood a particular grant applicant will have negative audit findings in the next year with 96% accuracy, and then quickly direct the Federal grant-maker to relevant data that helps them make better-informed decisions. This type of review and analysis used to take hours and days to complete and now takes just seconds.
Dovel is proud that this initial work has been recognized with a Government Innovation Award and excited about the capabilities in the pipeline over the next year. It’s clear that these technologies are essential building blocks to solving some of our nation’s biggest challenges, and will be particularly helpful in helping agencies achieve the goals and objectives laid out in the President’s Management Agenda. While we’re still in the early stages of applying these capabilities, we’re excited about the transformational impact they will have in delivering even greater value to the American People – in the grants community and beyond.