"AI does not fix broken processes. It accelerates good ones and exposes bad ones." — Jason Belford, Chief Information Security Officer for the City of Alexandria.
Jason Belford: staffing, risk, and why AI matters
Jason Belford, the City of Alexandria's CISO and a panelist at the Ivanti Public Sector Summit session "Harnessing AI and Automation for IT Service Management in State, Local Government and Education," frames the AI conversation around an immediate operational problem: understaffed public agencies. He told Government Technology Insider that "the public sector has struggled to recruit and retain top talent," and that governments compete not only for general IT professionals but for cybersecurity professionals, cloud engineers, AI specialists, and data professionals. In his words, tight budgets make it hard to match private-sector salaries, and in Northern Virginia the competition is particularly intense.
Belford warns that staffing shortages concentrate critical knowledge in fewer hands, increasing organizational risk and making documentation, knowledge sharing, and resilience more important. He draws a clear line between workforce constraints and service outcomes: delays in permitting, public safety, social services, infrastructure, and other functions can "reduce public confidence and create even more work as backlogs grow." His imperative is practical: modernize processes and introduce automation and AI "responsibly" so staff can focus on mission-critical tasks.
Where AI fits: routine, rules-based, and measurable tasks
Belford describes AI and automation as tools that "work best when they are applied to the right problems." He identifies the strongest candidates for automation as "routine, repetitive, and rules-based activities" supported by well-defined processes, clear policies, and accurate data. Specific examples he mentions include resident-facing and internal operations such as permitting and planning workflows, information security risk assessments, and "other administrative processes where automation can improve consistency, reduce processing time, and allow employees to focus on higher-value, more complex work."
He stresses process-first thinking: "I have always believed in focusing on people, process, data, and then technology. Technology should improve a well-functioning process, not make up for a poorly designed one." From Belford's perspective, AI should accelerate effective processes and expose weaknesses in poor ones — not mask them.
Trust, governance, and the limits of automation
Belford is explicit that trust in AI must be earned. "AI should not be trusted simply because it is AI," he said. He sets out a checklist of governance and security requirements that echo throughout his remarks: evaluate risks, validate performance, establish safeguards, and continuously monitor results. He adds that AI "should be held to the same standard" as other enterprise technologies.
On decision-making, Belford draws a cautionary distinction: AI can analyze information, identify patterns, recommend actions, and automate routine tasks, but "decisions that significantly impact residents should continue to include meaningful human oversight and accountability." He also calls for transparency about how AI is used — what data it relies on, its limitations, and how results are validated — so that AI-assisted decisions remain explainable and auditable at the appropriate risk level.
What attendees at the Ivanti Public Sector Summit will learn
At the Ivanti session, Belford says attendees should expect practical guidance rather than hype. He wants them to leave "with a practical framework for evaluating AI opportunities instead of feeling pressured to adopt AI simply because it is the latest technology trend." Panel topics will include identifying business processes suitable for AI, defining where human oversight must remain, and building governance to deploy AI responsibly. He will share "lessons learned, challenges we have faced, and approaches that have helped us move forward while maintaining security, privacy, and public trust."
He underscores that successful AI initiatives "begin long before an AI model is deployed" — with well-defined processes, quality data, effective governance, and clarity about desired outcomes. For Belford, AI is a means to better service delivery, not an end in itself.
What this means for state and local technologists, policymakers, and residents
- State and local technologists and security teams: Expect to prioritize governance, data quality, and process standardization before deploying AI. Belford points to information security risk assessments and administrative workflows as immediate areas for exploration.
- Policymakers and procurement leaders: Belford advises choosing partners that demonstrate "strong security, privacy, governance, and accountability practices" and investing in oversight mechanisms so AI systems "earn" trust through testing and monitoring.
- Residents and public-facing employees: Belford frames AI as a tool to reduce repetitive work and improve consistency, enabling employees to spend more time on complex issues requiring human judgment, empathy, and experience — while preserving human accountability for high-impact decisions.
Jason Belford's argument is pragmatic: constrained budgets and scarce talent make automation attractive, but effective AI in government depends first on sound processes, governance, and transparent oversight. As he told attendees, "When used thoughtfully, AI can help government improve consistency, increase efficiency, strengthen public trust, and allow employees to spend more time serving their communities." The question the summit will test is whether agencies can put the necessary process and governance groundwork in place before turning models loose on services that affect people's lives.
https://governmenttechnologyinsider.com/can-ai-enable-more-efficient-delivery-of-citizen-services/
