To Advance Missions, Treat Agency Data Like a Product

Perspectives

To Advance Missions, Treat Agency Data Like a Product

Generative AI demands a new approach to agency data

With the latest generation of generative artificial intelligence (AI) tools like Bard and ChatGPT, AI鈥檚 long-heralded potential to transform professions, industries, and entire aspects of society is being realized and felt more than ever before. From our perspective as experts and executives in the space where government and technology meet, it鈥檚 clearer than ever that, wielded in the right way, generative AI and the innovation ecosystems that surround it could revolutionize the way that government serves the nation, its people, and its communities.

But it鈥檚 also clear that to make real on that tremendous potential, federal agencies must fundamentally change the way they think about their data, and they must take a coordinated, whole-of-government approach to doing so.

The practice of allowing most government data to stay cloistered within system and mission siloes must end. Instead, it鈥檚 time that agencies begin treating their data like a product鈥攐ptimizing it for secure and convenient consumption so that other approved entities, like other federal agencies, research institutions, and state and local programs, can harvest mission-advancing insights and services from it using AI and other tools. Our company鈥檚 heritage of supporting federal missions since the World War II era has convinced us and our colleagues at 无忧传媒 that in order to secure a bright future for America, government must put the full power of trusted, high-quality data to work on the nation's greatest challenges. Treating data like a product is a key step toward doing so.

Perspectives

Powering Up Data-Enabled Missions Across Government

How federal agencies can fully leverage their data to advance mission outcomes. Read More

Perspectives

Critically, a product framework would incentivize agencies to clean up their data for easier use, refining it from a messy, decontextualized, raw material state into the streamlined fuel that gives AI, modern software applications, and other digital tools their power to produce better mission outcomes throughout government. By aligning on a share-first, productized approach to data sharing, agencies can boost each other鈥檚 missions and truly work in concert toward improving lives and strengthening communities across the nation.

It鈥檚 important to note that generative AI is just one timely example of the many, many technological tools and processes that could leverage better integrated and organized agency data for the advancement of missions and creation of government services. For example, sites like and illustrate how modern software development, open APIs, and cleanly integrated, high-quality data can combine to create commercial-level government service experiences for the American public.

In terms of the difference the insights and outcomes it fuels could make in citizens鈥� lives, federal mission data is obviously an extremely valuable asset. There are many obstacles to empowering agencies to access and unlock more of that value by treating their data like a product鈥攊ncluding security concerns, data quality issues, serious questions around privacy, and an array of other cultural, administrative, and technical challenges鈥攂ut they must be overcome considering the transformative benefits that would result.

How Missions Would Benefit from Improved Intra- and Interagency Data Exchange

When novel data sources are introduced to mission analyses, solutions to long-standing challenges can suddenly become clear. To better understand how clean, organized data products from other offices, programs, and agencies could significantly boost mission achievement, consider an example from finance. 鈥溾€� has long been an essential presence in the bag of tricks that hedge funds employ to stay ahead of the market. In one now-legendary instance, a fund used satellite images to count cars in big-box store parking lots to guide an investment strategy around major retailers that netted tens of millions of dollars.

In government, just as in finance, the threads of information that prove key to revealing the whole rich tapestry can come from diverse, far-flung sources. A simple step like to autofill their tax data by pulling it directly from the IRS can reduce barriers to higher education by making applying for aid a less time-consuming, stress-inducing process. By consulting enrollment data from a program with similar criteria, Pennsylvania can identify and new prescription assistance benefit for older state residents. By pulling in data from global climate models, the to predict and plan for climate change鈥檚 possible impacts on thousands of American military installations.

Ad hoc examples of improved mission outcomes deriving from government data sharing abound, but imagine what could be achieved with agency-enterprise and even whole-of-government approaches to enabling the safe, easy sharing of analysis-ready program data products. Complex federal missions overlap with each other in myriad ways, whether it鈥檚 offering different services to many of the same people鈥擴SDA food stamps and HUD housing assistance going to the same household, for example鈥攐r tackling the same problem from different angles鈥攁s in public health agencies trying to better understand a dangerous novel virus, while drug regulatory programs screen potential vaccines for safety, efficacy, and possible approval. With so much synergy, and the hot potential that generative AI and other established and emerging technologies bring to the table, the opportunities for improved government efficacy through productized intra- and interagency data sharing are near infinite.

Begin the Journey to Better Data Access

Federal agencies still have a long way to go on their data optimization journeys, but recent developments should ease the way. On the technology side, there are now open-source and commercial solutions that make productively sharing data within and between organizations easier than ever. Data mesh and associated solutions, for example, allow data to be shared without having to engage in the laborious, once-standard practice of moving it to a specially constructed central repository. In terms of policy and action, the noncritical barriers to sharing data can fall when lives are on the line, and the Biden Administration has to tackle what鈥檚 perhaps the most important concern around data sharing in a democratic society鈥攑reserving individual privacy.

Here are some ways that federal agency leaders can help keep momentum moving in the right direction:

With and generative AI knocking at the door, the case for data sharing is now strong enough that federal managers and leaders should replace questions of whether or not a particular dataset should be shared with questions of how can it be made observable, accessible, and explainable in a secure manner for the whole of government to use for broad societal benefit.

This is one of many areas where thinking about data through a product framework can really help. To be successful, a product must cater to the needs of all of the user audiences it hopes to attract. In this case, that means federal data products must not only be easy to consume for those coming from outside an agency or program to reuse its data鈥攖hey must also be responsive to the concerns of those from inside the agency who are sharing their data but want to maintain awareness of who is using it and for what. Equally essential to a given product鈥檚 success is its ability to operate without causing problems. A government data product that causes harm by exposing private or sensitive information to the wrong users will obviously fail, thus taking a product mindset requires such risks to be anticipated and mitigated.

Stay Up to Date on What鈥檚 Technically Possible

New technologies and approaches are bringing once colossal data challenges down to size. For example, worries around losing control of one鈥檚 program data can be a major obstacle to scaling a share-first mindset across an enterprise. Today you can go a long way toward alleviating such concerns by building automated usage tracking and governance into an enterprise data sharing platform, allowing those who are sharing their data to track how and by whom it is being used.

Data mesh frameworks are slaying another monster data challenge of yore: data migration. Copying everyone鈥檚 data and moving it to a central repository was difficult and cumbersome and鈥攖hank goodness鈥攊t鈥檚 no longer necessary. Data mesh means it鈥檚 possible to make data accessible without moving it around.聽

Team Up Mission Experts with Technical Experts

Even artificially intelligent technology must be pointed in the right direction to succeed. Getting the most value from your data requires teaming technologists and data scientists directly with mission experts to ensure that the right questions are being asked and the right challenges are being targeted in the most technically effective ways.

Generative AI is rapidly lowering technical barriers, allowing more and more non-experts to set AI to work on their toughest challenges. But even generative AI tools like large language models to the needs of complex subjects and missions.

The interrelated nature of federal missions means that the speed and efficiency with which information is shared between programs and agencies鈥攊f it鈥檚 shared at all鈥攃an make the difference between success and failure, and good citizen experiences and bad. Groundbreaking advances in generative AI will only make this more true moving forward, with untold possibilities hinging on how quickly agencies can clean up and open up more of their data. By adapting to think of their data as a resource to be offered as a product that is used for the benefit of the citizen, agencies can help achieve their own goals as well as those of other agencies and the nation as a whole.聽

Learn more about 无忧传媒鈥檚 services for civil government missions

www.boozallen.com/civil

Meet Our Experts

John Larson

Dan Tucker

Article

1 - 4 of 8

无忧传媒