One Bot Saves NSF Staff Roughly 25,000 Hours Per Year, CIO Says


And it’s only one small piece of the agency’s broader automation vision.

The last few years mark the early end of the National Science Foundation’s journey adopting robotic process automation, the buzzy emerging technology practice that involves deploying bots or digital assistants to essentially mimic humans performing repetitive, menial tasks.

“We have automations going on in the [information technology] shop to facilitate automated build and deploy of solutions. We have the HR shop, the financial shop—they're all administrative functions,” NSF Chief Information Officer Dorothy Aronson said Wednesday. “But the one I'm most excited about is actually a brainchild of someone who was in the program office, who is not a technician, and just had this idea for how to save time.”

During a virtual interview, Aronson briefed Nextgov on that implementation and her broader RPA vision. She and another NSF insider also shared a little advice for feds looking to launch their own automations.

Taking Away the Mundane

RPA encompasses automating business processes that typically require more manual work. 

Elaborating on the existing use case that’s really stuck with her, Aronson explained that the NSF plans thousands of meetings each year, through which it hosts heaps of invitees. A large percentage of people generally don’t respond to the initial invitations that the agency’s business system disseminates, but certain rules around public meetings require a reply.

According to Aronson, Clarissa Johnson, who now serves as an NSF IT specialist, was sparked by an idea for how to save time in parts of that meeting-planning federal administrative function.

“She came up with a logic for automating the nag notes—or the notes that continuously go out to people—and it's been a tremendous time-saver,” Aronson explained. “We’re estimating that 25,000 hours a year are saved by administrative staff all throughout the agency because these notes are being sent by the bot.” 

The CIO said it’s become a “great boon” inside NSF, and that she’s excited about the application for numerous reasons. In particular, Aronson was delighted to see how a personnel member in one particular segment of the organization, who didn’t have much prior technical training, was able to employ a technology-enabled solution that ultimately benefited the whole agency.

“By working as a partner with the IT shop, she learned a lot about how IT people think, so that partnership was really important in her personal growth,” Aronson said, adding “she learned more about her own strengths—and I look forward to seeing where she goes with her career.”

Researchers have, for years now, flagged the potential of emerging technologies like RPA to potentially displace humans from their jobs. Still, Aronson and many others argue the tools are generally meant to help elevate individuals’ work experiences. 

“My overall feeling about that is that the automation of things should take away work that is mundane, so that people can do work that's more interesting,” Aronson explained. “My personal feeling here is that if there's something that a person loves to do, then it doesn't fall into that category—and so you don't automate things that are not troublesome.” 

Further, the CIO believes the future is going to be very different from the present, and that down the line “what people are going to want to do are the things that are uniquely human.”

Pointing to an analogy she said she frequently turns to on this topic, Aronson explained that many years ago, people had ice boxes, and people (ice men, usually) would go around and chip pieces of ice off for each of the houses. “Then of course refrigerators came along and no one asked ‘well, what happened to the ice man?’’ Aronson noted, adding that she imagines many of those individuals may have had interest in refrigeration and moved to work at repairing such appliances. 

But now, there are self-repairing refrigerators.

“So, what does the refrigerator[person] do then? And so I always like to think that they go off and teach yoga. I mean, you know, the world is changing in such a way that you don't really know what's going to displace things. But you know that the world is moving forward,” Aronson said, noting that this is part of the reason humans should be prepared to adapt to continuous change. 

“I think that we are now recognizing in advance that careers are going to have to change over time. Whereas I think that when I was a kid, maybe I grew up thinking ‘you get a job and then you go up the chain of command in that job, and you just do that one thing,’” she said. “I don't think that's the future. I don't even think that's the present—but people are not really aware of that right now.”

Community’s One Key

Automation implementations vary across U.S. federal agencies but generally fall under CIOs’ purview. 

“My initial fantasy for RPA was that we would stand up a central program that would empower the citizen scientists at NSF to create their own solutions,” Aronson explained. 

Officials several years ago within the agency’s financial group had identified relatively immediate opportunities to automate processes, so they and others reviewed options to eventually choose a commercially available solution that made sense. 

“But what happened was the tools on the market right now are not yet easy enough for typical end-users to use by themselves,” Aronson said. “And this is where [Anju Anand] really came in.” 

Anand serves as an RPA Program Manager at NSF. At that time, she helped stand a core infrastructure and develop policies, procedures and forms that end-users could turn to when documenting their automation-aligned work. Agency insiders were looking to implement a distributed development approach that could be scaled across the agency, and also ensure the central IT shop was aware of all automation in place. 

Unfortunately though, Aronson said that while some NSF personnel were able to pick up the skills necessary to develop RPA to match their ideas, most don’t. 

“Even now, most people who have ideas and who have implemented solutions are really not doing it themselves, but they have an intermediary technical assistant to do that work,” Aronson said. “I'm still hopeful, though, that we will get to a point where end-users will be able to develop their own solutions—and I still believe that's on the marketplace. The marketplace has to come up with the tools that will be even simpler than the ones we've got now.”

The CIO added that she views the agency on a promising path toward fully embracing RPA.

“There's a very broad range of implementations for robotic process automation in production today, and so that's a success,” she added. “I think it's just that they still are relying on a technical intermediary to implement the solution.”

Reflecting on some of what she’s learned so far from these pursuits, Aronson noted that networking, relationships, and “building a community around” RPA at the federal level have proven crucial in the agency’s efforts to use this technology. She confirmed that Anand leads an RPA community of practice, or COP, directly within NSF—and participates as a member in the broader General Services Administration RPA COP, where others from across agencies come together to share their work. 

“Being part of the Federal COP, I am seeing more and more people using [RPA] and how it can bring benefits,” Anand told Nextgov Wednesday. “And agencies are willing to share their code with each other. So, that enables us to benefit from what one agency has done, and try to replicate it in the other agency with minimal effort. That’s where I’m learning.”

“If I was on the street, giving advice to someone who said ‘I want to start an RPA program,’ I'd say, ‘reach out to GSA and join the community of practice,” CIO Aronson further emphasized.

As for other learnings, she added that it’s important for those just starting to venture into automation to accept that there’s always going to be a bit of a risk that the solutions being pursued might not catch on.

But to her, it’s always worth that risk.

“That would be my advice: Whether this succeeds or fails, we had to experience this ourselves. It's a cultural change, and cultural change takes time,” Aronson said. “I think low-code solutions are the way of the future. So, if someone invests in a low-code solution like this, and it ends up not to succeed, they haven't lost millions and millions of dollars of investment. You know, you pay as you go. It's a good way of going.”