"This is a new frontier for everyone," Adarryl Roberts told reporters at DLA's Industry Collider Day on Sept. 9 in Alexandria, Va.
Adarryl Roberts on "digital employees"
DLA's chief information officer used that phrase to describe a deliberate shift: the agency is moving from traditional robotic process automation to what he calls agentic-AI—autonomous software actors that DLA will treat as "digital employees." Roberts framed the transition as cultural as much as technical, saying workers must become comfortable supervising and trusting outcomes produced by AI that runs 24/7 behind the scenes.
Roberts highlighted management questions that follow from this framing. "How are we managing our digital employees?" he asked, then spelled out the practical dilemmas: "How do you train a workforce to where everyone is essentially working as a supervisor because of the employment of AI with humans? Do you terminate an agent because of the performance?"
Unattended bots today: 185–190 digital workers and 300,000 hours saved
The autonomous tools are not hypothetical. Roberts said DLA currently operates "approximately 185 to 190 bots" and estimated that in 2025 those bots saved the agency about 300,000 hours of work. He added that roughly "90% to 95% of those are unattended bots," meaning they operate without direct human initiation.
Those bot-hours were applied to core DLA functions—getting food, fuel, medical supplies and clothing where they need to go—underscoring that the agency sees automation as part of the logistics backbone rather than a peripheral experiment.
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 scrambleGenAI platform and training: Gemini, ChatGPT, Grok, and "GenAI 101 training"
DLA provides defense workers with a platform that now includes Gemini and military-specific versions of ChatGPT and Grok for information that is sensitive but unclassified. Roberts described "GenAI 101 training" as the entry point for workers; the training is intended to familiarize staff with generative tools, allow interaction with results, and build trust in what those tools produce.
"I would look at generative AI as the concept of making the workforce and agency—from an operating model—comfortable with the results you'll get with AI," Roberts said. He then connected that comfort to the next step: "Now, we're shifting to what I would call strategic implementation in terms of those agentic agents to get us better results."
DLA AI Center of Excellence, zero-trust, and persona-based access
Institutional guardrails are already in place. DLA founded an AI Center of Excellence in 2024 to "ensure the safe, secure, and responsible integration of AI across the agency" and to oversee agent training. Complementing that effort, Roberts described a zero-trust cybersecurity framework as "foundational" to high-level agentic work.
On access control, the agency is establishing "persona-based access" so that AI agents receive permissions aligned with job descriptions. Roberts offered an example: a procurement specialist's digital agent should access procurement and financial data domains, and no more. "These AI agents should have a persona, just like a person," he said.
Data hygiene is part of the program. Roberts said DLA is working to create a "single source of truth"—a "gold standard" dataset that removes duplicity and error so data is usable not only for analytics and AI but for "future system sustainability."
What this means for procurement specialists, DLA managers, and security teams
- Procurement specialists: Digital agents will be given narrowly defined persona-based permissions tied to procurement and financial data domains; specialists will need to supervise agents that perform equivalent duties.
- DLA managers and supervisors: As Roberts noted, nearly every human employee may act as a supervisor to AI agents—raising questions about performance management and threshold criteria such as whether to "terminate an agent because of the performance."
- Security and engineering teams: The agency's zero-trust baseline and the AI Center of Excellence will be expected to certify agent training, govern access, and support the effort to produce a single, reliable dataset for agent consumption.
Lt. Gen. Mark Simerly: "AI is the new gunpowder" and the operational imperative
Director Lt. Gen. Mark Simerly, whose 32-month tenure concludes with a retirement on Oct. 2, framed the shift in operational terms. "I believe when it comes to military logistics, AI is the new gunpowder," he told reporters, adding that AI "can also be very productive" in addition to being potentially destructive.
Simerly said that recent conflicts informed the urgency of change. He cited the contested logistics in Ukraine and a seven-month-old war on Iran launched by President Donald Trump as drivers for accelerating DLA's transformation. He also referenced Operation Epic Fury by name when discussing lessons about alternate supply chains, precision targeting, and the need for adaptability.
Simerly described a philosophy of moving "at pace with the commercial world"—adopting capabilities quickly, building networks and infrastructure to manage agency data, and enabling interoperability between the services and industry to "operate end-to-end supply chain operations that are optimized."
As DLA prepares for a change in leadership—Air Force Lt. Gen. David Sanford is slated to replace Simerly—the agency is explicit about its phase: Roberts called the work "early in that journey." The combination of unattended bots already saving hundreds of thousands of human hours, new agentic systems on generative-AI platforms, persona-based permissions, and an institutional AI center creates an experiment in governance as much as technology. The next steps—how those agents are managed day-to-day, how performance and failure are judged, and how data is solidified into a "single source of truth"—are concrete tasks the agency has set for itself as it moves agents from pilot to production.
