Thread AI has signed a research agreement with the US Army to develop artificial intelligence tools for Fire Control Systems research using its Lemma orchestration platform.
The Cooperative Research and Development Agreement (CRADA) brings Thread AI together with the US Army DEVCOM Armaments Center. The collaboration will allow defence researchers to use the platform to analyse extensive technical documentation and evaluate how AI can support specialised weapons research.
Colin Bell, AI Strategy at Thread AI, explained that fire control research relies on decades of highly specialised technical knowledge that general-purpose AI models are not specifically trained to understand. He emphasised that the challenge extends beyond the underlying model to the infrastructure required to connect information sources, reason across technical records, and trace conclusions back to their original documentation.
Thread AI was co-founded by Angela McNeal, its CEO, alongside former Palantir AI product and engineering leaders. The company focuses on developing infrastructure that enables organisations to deploy AI systems in complex, highly regulated environments.
McNeal said that the growing challenge for critical operations is ensuring the surrounding infrastructure is ready for real-world deployment, with the necessary security, governance, and operational controls.
Applying AI to Complex Defence Research
Lemma serves as a composable infrastructure platform designed to orchestrate enterprise workflows. The DEVCOM Armaments Center selected the technology to help military scientists examine extensive archives related to weapons development.
The platform coordinates data pipelines and autonomous software agents, providing the infrastructure needed to support agent-based workflows at scale. Governance and reliability controls are intended to help ensure that these processes operate within defined requirements.
Traditionally, analysing specialised defence documentation has required researchers to spend considerable time reviewing historical technical papers, system records, and other complex materials manually. Thread AI and the Army aim to develop a dedicated application that can synthesise these information sources more efficiently.
The proposed system will help researchers move from an initial question to a relevant technical finding more quickly, while maintaining controlled access to sensitive information and adhering to strict security requirements.
Traceability is a central component of the initiative. Lemma is designed to link AI-generated conclusions to their underlying source materials, allowing researchers to verify findings against the original technical documentation. This approach could help address concerns about the reliability of AI-generated answers in environments where analytical accuracy and accountability are essential.
Bringing Governed AI Into Defence Operations
The agreement represents a step in Thread AI’s expansion into federal defence programmes, building on its work in regulated enterprise environments.
Government agencies face increasing pressure to adopt modern computational tools across research and operational disciplines. However, introducing AI into military environments requires more than technical capabilities. Systems must also satisfy stringent security policies, preserve operational control, and support dependable deployment.
McNeal said composable infrastructure can help the Department of Defense advance its AI-first objectives at a pace comparable to commercial technology development without sacrificing mission security or control.
Under the agreement, engineers from Thread AI and the DEVCOM Armaments Center will collaborate on building the research application using Lemma. The resulting tool is intended to provide governed workflow capabilities tailored to Fire Control Systems research across the centre’s facilities.
The project highlights the growing role of AI orchestration infrastructure in specialised government applications. Rather than relying solely on general-purpose models, organisations can use orchestration platforms to connect data, coordinate automated processes, and make AI-generated findings easier to verify.
For defence research, this combination of information synthesis, workflow automation, and source-level traceability could help researchers navigate large technical archives while maintaining the oversight required for sensitive work.
