Families in Michigan rely on the state’s Department of Health and Human Services to access critical benefits. But first they must demonstrate their eligibility for state assistance — a process that relies on a monolithic legacy app called MI Bridges.
However, the state lacked visibility into the app and found it difficult to predict or respond to application outages, which had a negative impact on the people the agency serves.
“It’s a pretty old application,” Sanjay Srivastava, a division director with the state’s Department of Technology Management and Budget, told StateTech in an interview during Splunk’s .conf19 conference in Las Vegas.
“We’re in the process of modernizing the application. Downtimes were pretty high, and we wanted to make sure the application was stable,” he said. “But we knew it would take time and funding to rewrite the application, and in the meantime, we couldn’t impact the downtime.”
Josh Scheurer, a system architect for the state, added that Michigan “wanted critical insights into the application.”
“We wanted more observability into the inner workings of the application, which, since it’s a monolithic legacy application, was hard to get,” he said. “We didn’t have any kind of centralized reporting to look at — for dashboards, for triaging — to see what was going on in the application, to try make it more stable or available to the users with less errors.”
Michigan Uses Data Analytics to Improve Application Performance
The state invested in the data analytics tool around a year and a half ago, at first running a six-month pilot with its HHS agency. State IT officials had three goals for the pilot program: First, they wanted to attain that central visibility they were lacking in the benefits eligibility application. Next, they also wanted to be able to track and improve compliance with server-level agreements for uptime and mean time to recovery. Finally, they wanted to simplify and automate auditing for compliance with data safety standards.
“One of the major achievements was, once we were resolving these issues, it was helping technical, as well as the business,” Srivastava said. “With the data that was coming out of the logs, we were able to show the business, ‘Here’s your data, this particular data is showing you your resource load.’ And the technical department can also look at the back end to pinpoint where issues are happening, so we can prevent them before they impact caseworkers.”
The state plans to build on its success with Splunk by applying the tool to other HHS applications, such as those that support child services and Medicaid. Then, the state may expand use into other agencies and departments.
“It’s gaining more visibility: People are seeing it, there’s more demos, more interest from seeing the gains that we’ve had in our experience,” Scheurer said.
Scheurer added that there may be opportunities in the future to use analytics beyond support for IT applications. For instance, he said, he thinks states will soon be using data analytics for use cases such as fraud detection, with analytics tools able to sniff out illegitimate claims for benefits.
“We’ve done some piloting around fraud detection — people who are applying for voter registration and are applying for benefits under the same IP address, but as different individuals,” Scheurer said. “It hasn’t gone beyond piloting, but the potential is there. Things are rapidly changing.”