Leveraging AI to Tackle Big Challenges in Intelligence, Software Development and Fraud Protection
How new tools are helping government analysts, coders and consumer watchdogs work faster and more effectively
Three Points to Remember
- An advanced AI/ML platform that processes more than 10 terabytes of digital media every day is reducing backlogs and aiding search-and-discovery for intelligence analysts.
- At the Federal Aviation Administration, a generative AI assistant for software development is automating many repetitive tasks for coders and raising quality and productivity.
- The Federal Trade Commission is deploying a generative AI system to create consumer-complaint summaries that speed information to fraud investigators and law enforcement.
Across government, one issue is consistent: making sense of overwhelming amounts of data. Intelligence analysts may face billions of files in search of one critical lead, software developers can get stuck in routine coding and testing, and consumer protection teams are confronted with millions of fraud reports.
But with Leidos’ mission expertise and artificial intelligence support, government agencies are fielding reliable, secure and transparent AI tools that help them with data processing and management as well as insights for more impactful decisions.
Transforming raw data into actionable intelligence at scale
For more than two decades, tactical U.S. military operations worldwide captured large quantities of digital media on devices ranging from phones to hard drives, in pursuit of critical intelligence. Those archives piled up faster than analysts could review them. Today, with legacy media mixed with mountains of data flowing in daily, the need to process at scale is a national security imperative.
To address this challenge, Leidos developed and implemented an advanced AI/ML-enabled digital media exploitation platform within the U.S. Intelligence Community. The platform processes more than 10 terabytes of data per day and reduces the time required to identify reporting from weeks or months to seconds or hours.
“We have enabled the Intelligence Community to fuse years and petabytes of priority historical collections into a single repository, making previously difficult-to-access data discoverable for various mission needs, including counterterrorism, emerging conflicts and border security,” said Matt Reardon, Leidos' senior director of intelligence analytics solutions.
We have enabled the Intelligence Community to fuse years and petabytes of priority historical collections into a single repository, making previously difficult-to-access data discoverable for various mission needs.
Matt Reardon
Leidos Senior Director of Intelligence Analytics Solutions
The platform has integrated AI-driven capabilities, including computer vision, voice transcription, facial similarity analysis and language translation, to extract text, audio, images and video from raw digital files.
Generative AI workflows organize triage reporting by relevance and relationship, freeing analysts to focus on making connections and insights from within the data. Optimized workflows can enable faster identification of networks, trafficking routes and illicit finance operations.
Today, analysts can discover and track suspected terror groups, smugglers and criminals, mapping their digital footprints in a fraction of the time previously dedicated to these activities.
“We continue to push the boundaries of integrating AI at scale, evolving from multi-modal machine learning classifiers to automated reporting powered by large language models,” said Ibrahim Shafi, the platform’s chief system architect and a Leidos tech fellow.
Freeing FAA’s software developers for innovation
While AI is assisting the hunt for threats to our national security, at the Federal Aviation Administration (FAA), software teams are taking advantage of AI to ship secure, high‑quality software that results in day‑to‑day improvements to the air traffic mission.
FAA’s software engineers build and maintain mission-critical systems from radar control to flight tracking — work that demands precision but often involves repetitive coding, debugging and testing. A generative-AI software development assistant, integrated directly into Leidos’ secure environment and tuned to aviation frameworks, helps them by automating routine code, documentation and test cases.
The assistant also provides instant debugging and flags security concerns, explains legacy code and suggests new mission functionalities.
In a pilot program, the assistant delivered a 62 percent productivity increase and a 27 percent improvement in defect resolution, shifting software developers’ time and focus from maintenance to improving safety, analytics and operational efficiency.
“Our pilot program demonstrated how AI technologies can bring substantial benefits to the software development lifecycle for air traffic programs, enhancing productivity and improving the quality of our delivered systems,” said Marcel Cote, Leidos’ senior director of transportation solutions. “Looking ahead, we are optimistic about the future as Leidos continues to incorporate these technologies to drive innovation and deliver even greater value to our customers.”
Protecting consumers with smarter tools
The American public benefits from safer and more efficient air traffic as a result of software improvements at the FAA, and AI is also helping to protect them from fraud.
A generative AI system that’s backed by more than 40 machine learning models now helps the Federal Trade Commission (FTC) take reports of consumer complaint calls, process mailed-in complaint letters and spot patterns in the complaints.
Through the Sentinel Network Services (SNS) program, which Leidos has been supporting since 2007, the agency handles more than 10 million consumer reports and 25 million interactions annually, with the AI models identifying common scam language, patterns and regional shifts in fraud activities with approximately 90 percent prediction accuracy to give law enforcement a faster read on trends.
At the FTC’s Consumer Response Center, the AI system transcribes and summarizes calls using an assistant that understands natural speech, reducing the reliance on live agents to manually record the interactions. Leidos’ generative AI solution also reads and organizes complaint letters that are often written by hand and in multiple languages, extracting key details, classifying them according to case type and generating clear summaries.
The AI assistant has cut processing time in half, slashed transcription errors and delivered 22 percent annual cost savings, and it is expected to help save the agency about $600,000 over five years. Together, the AI upgrades let FTC staff spend more time helping callers and get higher‑quality information to investigators faster.
With better, more accurate summaries and less need to conduct manual checks, auditors are seeing quality gains and investigators are receiving critical information sooner. Over the past five years, the SNS program has helped return $1.4 billion to consumers.
The bigger picture of AI-enhanced missions
Across intelligence, aviation and consumer protection, AI handles the heavy data lift while people steer the mission outcomes. Analysts weigh the connections that lead to more effective intelligence reporting, developers apply tools that deliver software more efficiently and fraud experts turn findings into help for the public.
Leidos builds capabilities that strengthen the human-and-AI partnership with reliable, secure and transparent systems that turn hard problems into measurable results.
MORE ON LEIDOS AI CAPABILITIES
