Insight by Carahsoft and Elastic

Preparing for advanced AI begins by getting your data ready

Chris Townsend, the vice president of U.S. public sector sales at Elastic, said for agencies to operationalize their data to drive better decisions, they need t...

Insights from the series, Innovation in Government.

June was the four year anniversary of the Federal Data Strategy and it’s clear how much progress agencies have made since OMB released that vision.

Every agency has a chief data officer and their own data strategies.

Governance and frameworks have been taking shape over the last few years and agencies now are starting to see the impact of all this work.

The Federal Data Strategy calls for agencies to use data to drive decisions, or operationalize their data.

When you add to that the advancements in artificial intelligence, machine learning and automation, having the right data at the right time and ensuring it’s accessible by the right people becomes more important than ever.

The IT Innovation Foundation found in a report from June that agencies need help to advance in the maturity model outlined in the 2021 update to the federal data strategy.

ITIF says agencies still need a lot of help to better understand which tools and techniques are most appropriate for them and their customers and how to prioritize data insights that are most important to their own missions.

Chris Townsend, the vice president of U.S. public sector sales at Elastic, said whether it’s effectively delivering health care process claims, performing fraud detection or better enabling the warfighter, agencies are trying to understand how to operationalize their data in the most efficient and most cost effective way possible.

“The scale at which agencies operate with data in multiple clouds and data on premise as well as  unstructured data and structured data, the challenge is how do they effectively operationalize that data across all those silos?” Townsend said on the Innovation in Government show sponsored by Carahsoft. “Some large agencies that have individual organizations within that agency, each of those sub agencies have their own data stores as well, and they’re not easily accessible. So being able to operationalize data at scale, and across all of these multiple silos is a real challenge.”

The solution to that challenge, Townsend said, is to take search to the data rather than bring the data to the search tools.

“The thinking for a long time is we have to bring all this data back to a common data store or centralized location like a data warehouse or a data lake. But that’s just not practical for a lot of our large complex public sector customers,” he said. “You hear the Defense Department talk a lot now about their data mesh strategy, which is the idea of being able to take the analytics in the search to the data, where the data resides, across all these different silos, and then bring just the relevant information back to a centralized location.”

CISA, DoD examples

Townsend said a good example of that approach is the Continuous Diagnostics and Mitigation (CDM) dashboard run by the Cybersecurity and Infrastructure Security Agency.

ECS won the $276 million CDM dashboard contract in 2019 to implement the Elastic technology.

Townsend said CISA is bringing together data from more than 100 disparate agencies and performing data analysis.

“The whole idea of operationalizing data at scale to support a mission outcome, whether it’s improving your cybersecurity posture or improving your threat hunting, the more data that you can get into that environment, access and normalize, the better your results are and the better the outcomes are,” he said. “In addition to that, if you’re trying to implement a better decision making process and analytics process across all of the Department of Defense, the Navy’s got its own data store, the Army’s got its own data store, the Air Force has their own, so how can you query all of that data in a common way in a common framework and be able to garner results across that entire environment.”

Townsend said another example is the DoD’s Joint Regional Security Stack (JRSS) effort that now includes, for the first time, the Army bringing together data at both the strategic and tactical environments through a unified architecture.

“You can’t replicate all that data and bring it back to a centralized location. You need to be able to search that data out at the edge, and then just bring the relevant information back in with JRSS. You’ve got these massive amounts of data flowing through those Joint Regional Security Stacks, and if you want to provide that dashboard and threat analytics, you can’t expect to bring all that data back to a centralized location because that just doesn’t work,” he said. “We’re talking a lot about security, but that applies to everything else to whether it’s artificial intelligence operations or in any data analytics or search function. We are increasingly doing more with government agencies around the executive order around improving customer experience. There’s a lot that we could do to improve interaction with government websites, and leveraging things like large language models on ChatGPT to make data more accessible to the citizenry.”

Data to fuel AI tools

Agencies still are dipping their toes into generative AI, but to prepare for a possible future they have to change the way they work with the data.

Townsend said generative AI tools search for information in the past, using keywords and providing meaning and context.

“I think everyone is obviously trying to understand how to best use that technology in a secure way,” he said. “Agencies recognize the tremendous potential benefit of being able to access data to do generative AI, and things like improving security posture, doing threat hunting, or providing better access to the citizen citizenry around the, executive order of improving customer experience. There are tons of applications and we’re just scratching the surface of these technologies.”

Townsend added the benefits of applying AI tools to securing data and protecting systems are another attractive aspect.

But to prepare for using AI in any of the potential areas, Townsend said agencies have to continue to get their data houses in order. He said while the federal data strategy has helped create some momentum, there are still things agencies can do.

“We’re starting to see a lot of convergence in building more cohesive agency-wide data strategies. I think we saw the use of data in pockets. If you had an operations group over here that was doing fraud detection, they may be indexing and using their data for something over here. The cybersecurity folks may be using data too and the customer experience folks may be using different data,” he said. “But now agencies are looking what should they be doing agency-wide and enterprise-wide in terms of their data strategy. What tools should they be consolidating? How are they indexing their data and whether they are duplicating their data and paying for multiple storage solutions. Are they paying for multiple tools to index the same data repeatedly? I think we’re seeing a lot of consolidation around data and seeing a lot of consolidation of the tool sets so that they can buy one tool set and be able to use multiple third-party solutions that can sit on top of a platform that can use that data in different ways.”

To learn more from the Innovation in Government series, click here.

Listen to the full show:

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.

Related Stories