Presented by Carahsoft
Federal agencies are moving rapidly beyond artificial intelligence experimentation and into large-scale operational deployment. Leaders across government and industry are using AI to modernize operations, improve citizen services, strengthen cybersecurity, combat fraud, and empower employees with new tools and insights. During the AI for Government Summit, executives from the Department of Energy, Centers for Medicare & Medicaid Services, the Central Intelligence Agency, and leading technology organizations shared how AI is transforming mission delivery while highlighting the challenges of data readiness, governance, security, workforce development, and trust. Together, these leaders painted a picture of a government increasingly focused on turning AI investments into measurable outcomes that improve services, enhance efficiency, and strengthen national competitiveness.
Zimmer explained that DOE's AI journey began with internally trained models that leveraged agency data. While useful, those early efforts were limited because they lacked access to broader information sources. By introducing web-grounded capabilities while keeping sensitive DOE information secure, the department dramatically expanded the value of its AI environment. Employees can now access relevant external information while maintaining the security and integrity of proprietary data.
One of the agency's most impactful developments has been Quanta, a data platform that enables the integration and analysis of massive datasets from multiple sources. Whether evaluating energy markets, grid modernization initiatives, or national energy strategies, leaders can now conduct sophisticated analyses that would have previously required extensive manual effort.
Zimmer noted that technology has not been the biggest challenge. Instead, changing organizational culture and encouraging data sharing across traditionally siloed programs has required significant effort. Convincing employees that data belongs to the enterprise rather than individual offices has been essential to unlocking the full value of AI.
The success of these initiatives has fueled rapid demand across the department and beyond. Other agencies are increasingly interested in DOE's approach, while executives throughout the organization are actively sharing prompts, insights, and best practices that continue to expand the value of AI across mission areas.
Key Takeaways
• DOE is integrating internal and external information sources to create more powerful AI tools.
• Large-scale data analytics are helping leaders make more informed energy and policy decisions.
• Cultural change and data sharing are critical enablers of successful AI adoption.
Adams observed growing enthusiasm across all levels of government as organizations seek practical examples of how AI can improve operations and citizen services. State and local governments, in particular, are accelerating adoption as they explore opportunities to improve efficiency, combat fraud, and support increasingly complex missions.
Central to AI readiness is ensuring that organizations have reliable, accessible, and governed data. Without strong data foundations, agencies will struggle to generate accurate insights or scale AI initiatives effectively. Infrastructure and security capabilities must also evolve to support the growing demands of AI workloads.
Equally important is workforce preparation. Adams stressed that AI should not replace subject matter expertise but rather enhance it. Human judgment remains essential to ensuring that AI-driven recommendations align with mission objectives and organizational priorities.
As AI capabilities continue to evolve, Adams believes agencies will increasingly focus on agentic workflows that automate routine tasks, support decision-making, and improve citizen-facing services. Organizations that invest in readiness today will be best positioned to realize those benefits in the future.
Key Takeaways
• AI readiness requires strong data governance, infrastructure, and workforce development.
• Human expertise remains essential to successful AI implementation.
• Agencies are increasingly focused on measurable mission outcomes from AI investments.
Oliphant noted that many organizations initially adopted AI to help employees draft documents, summarize information, or automate administrative tasks. While valuable, those use cases only scratch the surface of AI's potential. The greater opportunity lies in redesigning business processes around AI-enabled workflows that fundamentally improve how work gets done.
Data remains one of the largest obstacles to achieving this vision. Agencies often struggle with fragmented systems, outdated information, inconsistent governance, and legacy platforms that make integration difficult. Without clean, trusted, and accessible data, AI systems cannot deliver reliable results.
Another challenge involves information access and security. AI dramatically increases the ability to surface and analyze information, requiring agencies to carefully evaluate permissions, governance structures, and access controls. Leaders must ensure users can benefit from AI without exposing sensitive information.
Oliphant warned that agencies risk creating fragmented ecosystems filled with overlapping bots, agents, and tools if they fail to establish strategic frameworks. Organizations should instead focus on designing shared capabilities that support the highest-impact workflows while encouraging collaboration across teams.
Intentional design, she argued, will help agencies avoid duplication, reduce complexity, and maximize the long-term value of their AI investments.
Key Takeaways
• AI should be used to redesign workflows rather than simply automate existing tasks.
• Data quality and governance remain major barriers to enterprise AI adoption.
• Strategic planning helps prevent fragmented AI ecosystems and duplicate efforts.
One of CMS's most significant accomplishments has been its use of AI to combat fraud, waste, and abuse. Machine learning and advanced analytics enable the agency to identify suspicious claims patterns in near real time, allowing investigators to detect issues that previously might have taken years to uncover.
Brandt reported that these efforts have prevented approximately $2.1 billion in improper payments in just over a year. By automating analysis and identifying anomalies more quickly, AI is helping CMS focus resources where they can have the greatest impact.
The agency is also using AI to improve acquisition and contract management. Tools such as the CMS Labor Analysis Wizard compare proposed contracts against historical agreements, helping acquisition professionals identify cost-saving opportunities and negotiate more effectively.
Workforce education remains a major priority. CMS leaders are investing heavily in AI literacy programs designed to ensure employees understand how to responsibly use emerging technologies. The goal is for every employee to develop a foundational understanding of AI and its potential applications.
Looking ahead, Brandt sees significant opportunities to improve beneficiary experiences through AI-enabled applications that integrate health information, claims data, and personal health metrics to support better care decisions and improve health outcomes.
Key Takeaways
• AI has helped CMS prevent approximately $2.1 billion in improper payments.
• Advanced analytics are improving acquisition decisions and operational efficiency.
• Workforce education and beneficiary trust remain essential to long-term success.
Rollins explained that many agencies have successfully completed initial pilots and demonstrations but now face the challenge of operationalizing AI across the enterprise. Achieving that goal requires more than access to powerful models. Organizations must establish the infrastructure, governance, security controls, and data management practices necessary to support sustainable adoption.
Trust remains one of the most important considerations. Agencies must be confident that AI systems are producing accurate, reliable, and explainable results before integrating them into mission-critical workflows. Strong governance and oversight mechanisms help ensure those outcomes.
Workforce readiness is equally important. Employees need the knowledge and skills required to effectively leverage AI tools while understanding the limitations and risks associated with emerging technologies. Agencies that invest in training will be better positioned to maximize the value of their AI investments.
Rollins also highlighted the need for scalable architectures capable of supporting future innovation. As AI workloads grow more sophisticated, agencies will require infrastructure that can evolve alongside mission requirements while maintaining security and performance.
The organizations that succeed in the AI era, Rollins suggested, will be those that build trusted foundations today while remaining flexible enough to adapt to tomorrow's opportunities.
Key Takeaways
• Infrastructure is essential for scaling AI from pilot projects to enterprise operations.
• Trust, governance, and explainability are critical for mission adoption.
• Workforce readiness and flexible architectures support long-term AI success.
Over the past several years, the CIA has invested heavily in secure AI platforms, enterprise chatbot capabilities, and model-serving environments that allow advanced AI technologies to operate within classified environments. Those investments have established a foundation for broader adoption throughout the agency.
Today, the focus has shifted toward putting AI directly into the hands of officers who execute the mission every day. The agency encourages innovation at the operational edge while identifying successful solutions that can be scaled across the enterprise.
Soong emphasized that technology alone is not enough. Building trust in AI requires transparency, explainability, and human oversight. Because intelligence missions demand accuracy and accountability, users must understand how AI systems reach conclusions and remain confident in their outputs.
The CIA is also modernizing procurement and acquisition processes to accelerate access to emerging technologies while maintaining appropriate oversight and security standards. Partnerships with industry, research institutions, and venture organizations play an important role in identifying promising innovations.
As AI becomes more deeply embedded in intelligence operations, Soong believes workforce readiness and trust will determine the speed and success of adoption.
Key Takeaways
• The CIA is scaling AI from experimentation to enterprise mission support.
• Trust, explainability, and workforce readiness are critical adoption factors.
• Modernized procurement helps accelerate access to emerging technologies.
Lee described a rapidly evolving threat landscape in which cybercriminals use AI to automate attacks, mimic legitimate behavior, and identify vulnerabilities at unprecedented speed. This evolution requires equally sophisticated defensive capabilities.
Traditional approaches that focused on identifying malicious files are no longer sufficient. Behavioral analytics now play a critical role in identifying unusual activity, unauthorized access, and suspicious behavior that may indicate a compromise.
Lee believes AI systems themselves should be treated as critical infrastructure. Agencies must secure not only the data that feeds AI models but also the models themselves to ensure they have not been manipulated or compromised.
As organizations automate more decisions, governance becomes increasingly important. Agencies must determine where automation is appropriate, understand potential mission impacts, and maintain accountability for outcomes.
Ultimately, Lee sees cybersecurity as an enabler of AI adoption. Organizations that establish strong security foundations will be able to move faster and scale AI more confidently.
• AI is transforming both cyber defense and cyber threats.
• Behavioral analytics are becoming essential for identifying sophisticated attacks.
• AI systems should be protected as critical infrastructure.
Hayes noted that government organizations often operate in highly fragmented data environments. AI success depends on the ability to bring together structured and unstructured data while creating context that makes information accessible and useful.
Hybrid architectures are becoming increasingly common. Agencies are combining cloud environments, on-premises infrastructure, and edge capabilities to balance flexibility, performance, and cost.
Security remains a critical consideration throughout the AI lifecycle. Organizations must maintain consistent security policies and protections across all stages of data ingestion, processing, storage, and inference.
Looking forward, Hayes sees edge inference becoming one of the most important trends in government AI. Delivering insights directly to operational users will require secure and efficient movement of data throughout the enterprise.
Key Takeaways
• Strong data foundations are essential for successful AI adoption.
• Hybrid environments are becoming the dominant deployment model.
• Edge-based AI capabilities will drive future innovation across government.
Clay explained that threat actors are embracing agentic AI because it enables them to automate many aspects of cybercrime. As a result, organizations face more sophisticated attacks delivered at greater speed and volume than ever before.
Fortunately, defenders are also benefiting from AI advancements. Modern security platforms can now predict potential attack paths, continuously assess organizational risk, and recommend mitigation strategies before vulnerabilities are exploited.
AI is enabling organizations to move beyond reactive security and toward predictive defense. By analyzing vast amounts of information, security teams can identify weaknesses, prioritize resources, and respond more effectively to emerging threats.
Despite these advances, Clay emphasized the importance of maintaining human oversight. AI can accelerate analysis and automate tasks, but security leaders must remain accountable for critical decisions and outcomes.
The future of cybersecurity will increasingly depend on how effectively organizations combine AI-powered automation with human expertise.
Key Takeaways
• Cybercriminals are rapidly adopting AI to automate and scale attacks."
• Predictive security capabilities are improving threat detection and prevention.
• Human oversight remains essential even as security operations become more automated.