PACOM Tests Mission-Specific AI Agents for Intelligence and Targeting

The U.S. Indo-Pacific Command (USINDOPACOM) and its Joint Intelligence Operations Center (JIOC) are exploring how mission-specific AI agents and small language models could support intelligence operations, accelerate analysis, and reduce targeting delays. The effort is still in an experimentation and assessment phase, rather than a publicly identified acquisition. At TechNet Indo-Pacific 2026 in Honolulu, Scott Brunstetter, chief artificial intelligence officer for the J2/JIOC at USINDOPACOM, described the challenge: defense organizations may have multiple tools and multiple mission gaps, but no single model is suited to every task. As he put it, “There’s not one capability and one model to rule them all.”

The command’s interest in smaller, tailored models reflects practical constraints in operational environments. Brunstetter said small language models can address specific tasks while requiring less compute and power than larger frontier models. The intended approach is to match the right data and AI capability to the particular decision or mission, rather than assume one general-purpose system can serve all intelligence needs. For contractors, this points to potential relevance for power-efficient AI, deployable processing, edge computing, and tools that can be adapted to defined operational workflows. It does not, however, confirm a funded requirement or an active competition.

A central technical challenge is keeping AI-enabled capabilities useful in denied, degraded, intermittent, and low-bandwidth (DDIL) conditions. JIOC personnel are expected to conduct rapid experimentation to assess whether agents and other AI tools can continue operating under these constraints. Brunstetter said the next experimentation event was anticipated “in the next couple of months or six months,” indicating an approximate planning window rather than a firm procurement deadline. The stated operational goal is to ensure personnel can use relevant capabilities regardless of location, including when communications and access to centralized computing resources are constrained.

Data interoperability is another key element of the concept. Brig. Chris Alder, Vice Director for Intelligence, Australian Army, highlighted the potential operational payoff of common data standards: “And when you get to that common data standard, your latency for targeting reduces by minutes into seconds, which is where we need to be.” That emphasis links AI adoption to data integration and workflow design, not just model performance. Companies can use these articulated needs to assess whether their offerings support secure information exchange, tailored AI workflows, and tactical-edge operations, while treating industry discussions and future experimentation as possible engagement avenues—not as evidence of an open procurement.

• USINDOPACOM and JIOC are exploring mission-specific AI agents to support intelligence analysis and operational decisions.
• Small language models are under consideration for specialized tasks that require less compute and power than larger models.
• JIOC plans rapid experimentation to evaluate AI capabilities in DDIL environments; the source describes a possible event in the next few months or roughly six months, not a confirmed deadline.
• Operational needs include deployable processing and the ability for personnel to use AI tools regardless of location or communications conditions.
• Common data standards and interoperable integration are intended to improve information sharing and reduce targeting latency; Brig. Chris Alder said standardized data could move latency “from minutes into seconds.”
• Contractors with edge computing, power-efficient AI, data integration, or mission-tailored AI workflows may find their capabilities relevant to future discussions and experimentation.
• The reporting identifies no solicitation, contract award, budget, procurement contact, or procurement timeline; this is an emerging capability signal, not an active bid opportunity.
• Vendors should map existing products to specific intelligence and tactical-edge use cases and monitor future command experimentation and industry engagement for more concrete requirements.