“Agencies are trying to deal with all of the different options for AI out there, and how to govern, secure and control the models, as well as users,” said Scott Waite, Strategic Capture Director with Nutanix, drawing a through-line that kept returning at the August 26–27 Intelligence and National Security Summit: federal missions must run faster than the systems designed them to support.
The summit and the new tempo
More than 2,000 intelligence, defense, and national security professionals convened at the 2026 AFCEA and INSA Intelligence and National Security Summit to grapple with a straightforward but urgent fact: AI is accelerating operations, data volumes are exploding, adversaries are moving more quickly, and intelligence that once had value for hours may now need to reach decision makers within minutes or seconds. Those combined pressures are shifting speed from a competitive advantage to a requirement, participants agreed.
AI moves from experimentation to enterprise
Speakers at the Summit described a transition in federal AI posture. Where many agencies had previously treated large language models and other AI tools as experiments, those capabilities are now being embedded into “everyday mission operations.” With that adoption, however, came a familiar warning: deployment has outpaced governance and security. As Waite put it, agencies have stood up LLMs but “do not necessarily have the governance and security frameworks needed to manage them effectively.”
Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scrambleTrust, data, and the human role
Delegates stressed that greater AI use raises a core operational question: can agencies trust what AI produces? Summit discussion highlighted the difficulty of validating AI-generated results in mission-critical environments where accuracy and accountability matter. Human expertise remains essential, yet the volume and speed of incoming data make fully manual analysis impractical. Waite summarized the problem bluntly: agencies increasingly face data that is “humanly impossible to sort through.”
Those data volumes come from imagery, signals intelligence, and an expanding array of sensors, the Summit noted. AI can help identify patterns, anomalies, and potential signals buried in that stream, but the work is complicated by where data resides and who can access it across environments and security levels. The question the Summit repeatedly returned to was not whether AI can accelerate analysis but how to ensure its outputs are reliable and available to the right people at the right security level.
Intelligence sharing and shrinking timelines
Breaking through longstanding information silos took on new urgency in the face of faster threat activity. Summit participants warned that adversaries can exploit vulnerabilities, adapt tactics, and use AI to accelerate their own operations—shortening the window in which intelligence remains actionable. “The time frame to share that information to have it be valuable and effective just continues to shrink,” Waite said. In some scenarios, the relevant timeline is no longer days or hours, but minutes or seconds, and the ability to move actionable insights across organizations determines effectiveness.
Acquisition, integration, and operational sustainability
Faster acquisition was offered as part of the solution set, with agencies exploring alternatives to traditional procurement models to reduce the distance from capability identification to mission use. Approaches such as Other Transaction Authorities and Commercial Solutions Openings were highlighted as increasingly important when working with commercial and nontraditional technology providers. Yet Summit speakers cautioned that speed in procurement cannot substitute for the work of integrating, securing, governing, and operating new tools within complex federal environments: rapid buys still require operational sustainability.
Operationalizing AI: agents, platforms, and repeatable processes
AI itself is evolving, moving beyond basic agent loops toward more sophisticated agentic and multi-agent workflows. Waite noted that leading-edge AI is already advancing into multi-agent platforms and graph engineering—capabilities that may outpace an agency’s ability to evaluate, deploy, and govern them. The Summit framed a clear requirement: agencies need repeatable processes to evaluate, deploy, govern, and evolve AI so it “is available to everybody that needs it, with all the right controls around it.”
What this means for technologists, procurement leaders, and analysts
- Technologists and security teams should prioritize building governance and security frameworks that match actual deployments, since many LLMs are already in use without comprehensive controls.
- Procurement leaders must balance speed with sustainment—using authorities such as Other Transaction Authorities and Commercial Solutions Openings to accelerate delivery while ensuring integration, governance, and operational viability.
- Analysts and decision makers will need tools that surface the most relevant signals rapidly, preserving human expertise for validation and judgment even as AI handles bulk pattern recognition.
The summit’s collective judgment was clear: data will grow, adversaries will adapt, and the available time to turn information into action will continue to shrink. Success, participants argued, depends on pairing speed with the right guardrails—establishing governance alongside AI adoption, building infrastructure to accommodate evolving technologies and data volumes, improving information sharing across boundaries, and rethinking processes so proven capabilities reach the mission faster.
