Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.
The Trump administration has laid out a roadmap for artificial intelligence and predictive analytics in government. But in order to get those plans off the ground, agencies need to get a better handle on their data.
The White House will soon release a final version of its Federal Data Strategy, setting goals for agency chief data officers over the next year. But data experts say managing culture change remains one of their biggest challenges.
Tim Persons, the Government Accountability Office’s chief scientist, said agencies hiring data scientists could be a game-changer for how they carry out their mission, but he said it’ll take a different management approach.
“I think culture is the main frictional point on this, but I think we can seize that. There are certainly challenges — data are not often easy. We have often lots of it, but it’s not in the right form. So I don’t want to soft-pedal the idea that there are indeed challenges, but I think the opportunities and the ROI on this particular activity strategically is incredible,” Persons said Thursday on a panel hosted by GoveExec and Defense One.
In some cases, DoD has found some pockets of excellence. The Defense Logistics Agency, for example, has used robotic process automation, or RPA, to get a better handle on its supply chain data.
But Michael Conlin, the Defense Department’s chief data officer, said his job is to bring a data-centric culture to all of DoD, and tools like RPA can only get the Pentagon so far.
“People want to talk about technology and really cool tools, and there are wonderful tools … but if we don’t start from a cultural change, we’re not going to get where we need to go. The Department of Defense does not have a culture of data-centric decision making,” Conlin said.
One of the biggest opportunities to bring that data culture to DoD is modernizing its Defense Readiness Reporting System (DRRS).
The agency stood up DRRS in the early 2000s to get data-driven insights on its ability to carry out the National Security Defense Strategy. But Veronica Daigle, the assistant secretary of defense for readiness, said the time has come to give DRRS an overhaul.
“DRRS is very good at giving you a current snapshot, but it really wasn’t designed with that idea of predictive analysis, and that is definitely something that we know we need to do,” Daigle said. “And that’s also important just for planning purposes as we look forward: What are the forces that we’re going to need? How ready do we need them to be?”
Congress, as part of the 2019 defense spending bill, passed legislation that requires DoD to consolidate its readiness reporting systems.
Daigle said the overhaul will give DoD an opportunity to make the DRRS interface easier for users who input information, while Conlin said it’s an opportunity to make data-driven decisions at an enterprise level.
“Our challenge jointly with DRRS and readiness reporting — my challenge broadly is to enable decisions on a department-wide basis … There’s no sense in optimizing the Air Force and neglecting the Army, the Navy, the Marine Corps, the back office [and] the logistics functions,” Conlin said. “When you optimize individual components of an ecosystem, by definition, you will break the ecosystem.”
‘How do we do better?’ Chief data officers see hiring challenges persist
For all the challenges in standing up a data-centric culture in the DoD, Daigle said a lack of data isn’t one of them, but “trying to figure out what is the most meaningful data that I should be looking at to answer a particular question.”
The Foundations for Evidence-Based Policymaking Act tasks CDOs with using agency data to predict future outcomes and then measure their confidence in those predictions, but Conlin said the department still suffers from data quality issues.
“Part of our challenge has been [that] there are still human beings tapping on keyboards to put data into systems, instead of having it being generated automatically in the field through sensors, scanners and platforms. And that’s really what you want to be doing,” Conlin said.
To get a sense of just how much data DoD has, its recent digital modernization strategy showed the agency has more than 10,000 operational systems.
“Every one of those systems was put in place by a different team of people, meeting a different set of requirements, with a different budget, for different customers, in a different place over a time scale of 60 years of IT. It’s like working in an archaeological dig,” Conlin said.