The U.S. Department of Housing and Urban Development significantly overstated how much time its software bots would save workers, according to an inspector general report.
In 2018, HUD began its program for robotic process automation — technology that allows users to build software bots that can take over tedious and time-consuming tasks for humans, such as copying and pasting data, filling in forms or responding to emails.
HUD reported that 14 bots were being used that were saving 13,644 labor hours each year as of March 2022. The inspector general report found that only seven distinct bots were used, saving 1,015 labor hours or 7% of the goal.
“We found that many of HUD’s bots ... ultimately provided minimal business value,” the report said.
Federal agencies, through RPA and even AI, have been experimenting with technology to automate government work as a way to lower costs, free up workers and reduce errors for business processes and the millions of claims that come through public programs each year, whether at the Internal Revenue Service, the Social Security Administration or the Center for Medicaid and Medicare. Success has been mixed, however, because the type of work, funding and staffing varies between agencies, technology experts said at a recent panel on AI in government.
The General Services Administration, for examples, uses a bot to survey customers for building leases that has created an average of 260 hours of capacity per year per year. Before, more than 60 employees did that surveying manually.
RELATED
At HUD, results with automation floundered, partially as a result of the department’s decision not to renew its RPA services contract. Instead, the agency is supporting existing bots with limited contractor help and with a staff that it says is limitedly trained and unable to fix bots if they stop working or need tweaks.
Officials said staffing was their greatest challenge, as only two employees were assigned to manage RPA among having other roles. Adequate training would help, they said, but would still not be enough to ensure the agency could independently support its RPA program.
RELATED
The Office of the Chief Information Officer also did not monitor the bots, so reported results were inaccurate, the IG said.
As a result, costs were “excessive” compared to the minimal results achieved, the watchdog revealed.
In one instance, HUD allegedly spent at least $104,800 for a contractor to develop a bot that ideally would’ve saved 80 labor hours a year by sending notification emails to grant recipients. After a year, officials deemed the bot “impractical” and not any more useful than the standard Microsoft mail merge function that had been in use before.
In a memo on Feb. 1 responding to the inspector general, HUD officials said that since the RPA program started, agency CIO Beth Niblock and her team have made changes.
When asked about how it might address staffing needs to support RPA, the agency offered no further details.
Molly Weisner is a staff reporter for Federal Times where she covers labor, policy and contracting pertaining to the government workforce. She made previous stops at USA Today and McClatchy as a digital producer, and worked at The New York Times as a copy editor. Molly majored in journalism at the University of North Carolina at Chapel Hill.
