The US Air Force is turning to AI to tackle a major bottleneck in evaluating new systems: test data analysis.
To address this issue, the service has awarded San Francisco-based AI startup Unstructured a $2-million contract to build software that organizes complex testing datasets for rapid AI analysis.
Unstructured will expand multimodal data pipelines, allowing airmen to apply AI tools across fragmented testing data sources, including technical documents, imagery, audio, video, and telemetry streams.
The platform automates data processing, enabling personnel to query testing results through AI assistants instead of conducting time-intensive manual reviews.
By expanding AI access to test data, the program aims to shorten evaluation cycles and speed new capabilities to operational units.
The contract supports the Air Force Test Center modernization through the Next Generation Data Ecosystem (NGDE), an effort designed to connect data, tools, and teams across test programs.
“The game changer will be artificial intelligence as NGDE is actively engaged with AI projects,” said John Volk, the center’s acting chief technology officer and chief data officer.
He noted that AI will be critical to improving efficiency as workforce levels are unlikely to increase.
The award follows an earlier Department of the Air Force Digital Transformation Office contract selecting Unstructured to build a vendor-agnostic data foundation designed to scale AI across Air Force missions.
Unstructured CEO Brian Raymond said the effort will significantly compress analysis timelines: “These capabilities will allow Airmen to engage with vast amounts of technical data in seconds — data that previously required weeks or months to analyze.”
“By advancing multimodal data pipelines and building rigorous evaluation frameworks, we are helping the Air Force accelerate test cycles, improve confidence in AI systems, and deliver mission-critical capabilities to the warfighter faster.”
