Leidos and OpenAI bring AI deeper into federal operations
Leidos and OpenAI are teaming up to integrate generative and agentic AI into the day-to-day work of government agencies. The focus: digital modernization, health services, national security and infrastructure, and defense-core pillars of Leidos' NorthStar 2030 strategy.
The partnership centers on secure deployments that protect agency and customer data while delivering measurable efficiency. This isn't just a pilot push. The goal is real adoption inside core workflows.
What this means for operations leaders
- Shift from experiments to production: AI agents embedded in existing systems, not side projects.
- Faster cycle times: drafting, summarizing, and triaging across large document and data flows.
- Decision support: AI-assisted assessments that surface risks, anomalies, and recommended actions.
- Stronger service delivery: automation that shortens queues, reduces rework, and tightens SLAs.
Where AI will show up first
- Threat and risk assessments: AI-generated briefs, red flags, and source cross-checks.
- Supply chain monitoring: vendor screening, disruption alerts, and compliance checks.
- Media integrity and deepfake detection: automated screening before escalation to analysts.
- Case intake and triage: routing, summarization, and response drafting for high-volume channels.
Leidos is also rolling this tech into its own operations. Thousands of employees are using ChatGPT and the OpenAI API Platform to speed internal work, accelerate product development, and build custom agent workflows. Expect those internal gains to translate into faster delivery for customers.
Security and trust take center stage
The companies emphasize protected configurations and mission relevance as adoption principles. For federal teams, that aligns with current risk frameworks and governance expectations.
- Data protection: strict access controls, retention policies, and audit trails.
- Human-in-the-loop: keep oversight on critical decisions and escalations.
- Governance: document model use, testing, and performance thresholds across use cases.
For reference, see the NIST AI Risk Management Framework for a common approach to governance and controls. NIST AI RMF
How to move from interest to impact
- Pick 2-3 high-volume processes with clear metrics (backlog, handle time, accuracy). Ship a 60-90 day pilot.
- Set guardrails early: data classifications, red-teaming, policy for sensitive content, escalation paths.
- Integrate with what you already use: case systems, document stores, messaging, and RPA.
- Measure what matters: SLA adherence, rework reduction, cost-to-serve, and user satisfaction.
- Plan for change management: training, role clarity, and feedback loops for frontline teams.
Questions to ask your vendors
- How is data isolated and logged? What are the retention defaults and options?
- What model updates are planned and how are they validated before rollout?
- Can we get evaluation results on our data and tasks, not just general benchmarks?
- What deployment options exist (GovCloud, private endpoints, on-prem patterns)?
- How are incidents reported, and what's the remediation timeline?
About the companies
Leidos is a technology and solutions provider to government and commercial customers. Headquartered in Reston, Virginia, the company employs about 47,000 people and reported approximately $16.7 billion in revenue for the fiscal year ended January 3, 2025. Learn more at leidos.com.
OpenAI is an AI research and deployment company focused on ensuring advanced AI benefits everyone.
Level up your team's AI capability
If you're building an AI rollout plan for operations roles, these resources can help:
Bottom line: this partnership signals that AI is moving into the core of federal operations. Start small, measure hard, and build the governance now so scaling later is easier-and safer.
Your membership also unlocks:
