The Defense Logistics Agency’s deputy chief financial officer is exploring the power of artificial intelligence to accelerate the agency’s path to a clean financial audit.
“While we are in the early stages of using AI in finance, we believe there is a lot of potential. We have already seen success with technologies such as bots and robotic process automation. We’re going to build it with AI,” Sean Lennon said.
AI can transform DLA progress by loading policies, data, process documents, and more into large, searchable language models for easy and rapid retrieval. The goal is to input data from DLA’s business systems and use AI to detect errors, generate insights, and recommend solutions to improve data quality and financial reporting, Lennon added. I did.
With AI, employees can approve AI-suggested solutions and make corrections with robotic process automation.
“We are also looking at using AI to match DLA inventory in our warehouse management system with our financial records,” Lennon continued. “Many people are now manually reviewing error transactions, trying to figure out what went wrong, why, and possible solutions. There are too many for us to keep up with.”
According to the National Defense Authorization Act of 2024, the Department of Defense must submit a clean audit opinion by fiscal year 2028. The 2025 draft NDAA directs the Secretary of Defense to report to Congress on the department’s use of AI to accomplish that goal.
Lennon has volunteered to lead the Department of Defense Financial Management Task Force to determine how DLA, military services, and other defense agencies can incorporate AI for auditing purposes while protecting DoD data. I did. The group is studying use cases for industry and other federal agencies, including Defense Department agencies such as the Defense Financial Accounting Agency, which has already established governance for the use of AI in efforts such as fraud detection. I’m learning from experience.
The Government Accountability Office’s chief scientist recently taught John Lennon and participants in the Public Service AI Federal Leadership Program how to use large-scale language models to search and extract insights from a century of audit reports. I explained about it.
“The Department of Defense could do something similar. For example, if they want to know the history of the Joint Strike Fighter and all the audits that have been done, they can provide an overview and recommendations, as well as implementation status. If you want it, the AI will get it all to you right away,” Lennon said.
While each Department of Defense agency will likely use AI for audits in different ways, Lennon said the group is helping everyone benefit from the progress and setbacks of others. said. The Navy commander added that one agency’s ideas may appeal to other agencies as well. Jonathan Henson is an executive officer of DLA Finance.
“If we have a solution that works for different institutions, we might be able to pool resources and fund something. And everything that we’ve learned from a data tagging and security perspective will help us “Everyone can learn from it,” he said.
DLA leaders also hope to learn from other federal agencies, such as the State Department. The State Department has been implementing AI into its internal networks over the past year and will soon demonstrate DLA in action.
However, using AI to store and screen data is risky, especially when shared outside government agencies, as the information can end up in the wrong hands. said Lennon. GAO mitigated that risk by disconnecting the environment from the Internet.
“But it reduces some of your capabilities. So even if the data we’re collecting isn’t sensitive, these are trade-offs that we need to make,” he said. .
DLA Finance’s experience has been limited so far, as most large-scale language models for AI are web-based and not specifically designed for use by the Department of Defense. When the team gained temporary access to Microsoft Azure OpenAI, they used it to load large documents, such as the 1,500-page Department of Defense Financial Management Regulations, and perform data summaries on specific topics.
The team is currently working with DLA Information Operations to responsibly deploy AI tools and fund use cases.
“We need to decide how we prioritize what we do with AI and establish processes for approval and testing of AI, which is everything that comes with new technology,” Lennon said. he said.
Whether DLA needs to create employee positions specializing in AI or train existing employees in skills such as rapid engineering is another consideration.
While job security often comes to employees’ minds when it comes to new technology, Henson said AI can help employees spend less time on manual tasks such as managing spreadsheets and more on data analysis and decision-making. said they can spend more time doing meaningful work.
Lennon added that employees need to start thinking about the benefits of AI and receive training to better understand its potential.
“In support of our new strategic plan, we are adjusting employee expectations around how AI can be leveraged to more efficiently and effectively accomplish missions across the agency, not just in finance and audit. “And we want to increase our goals,” he said.
