In the landscape of modern municipal management, the integration of artificial intelligence is no longer a futuristic aspiration—it is an operational reality. As cities across the United States scramble to modernize aging infrastructure and bridge the gap between limited staff resources and increasing public demand, Raleigh, North Carolina, has emerged as a vanguard. By treating AI not as a "magic button" but as a new employee in need of mentorship, the city is rewriting the playbook for government digital transformation.
Under the guidance of Chief Information Officer Mark Wittenburg, Raleigh has become the first municipality in the U.S. to deploy ServiceNow’s L1 AI Specialist product. This strategic move, which began earlier this spring, marks a significant shift in how local governments handle internal administrative burdens, moving from manual, labor-intensive workflows to highly automated, agentic systems.
The Human-Centric Approach to Machine Learning
CIO Mark Wittenburg, who arrived in Raleigh three years ago with decades of IT experience from Tempe, Arizona, possesses a clear philosophy: AI implementation is a process of trust-building.
"You don’t bring somebody on to the help desk and then immediately give them admin permissions," Wittenburg explained in an interview with Smart Cities Dive. "You slowly build trust over time."
This "onboarding" analogy defines Raleigh’s strategy. Rather than deploying black-box algorithms across the entire municipal enterprise, the city has opted for a measured, tiered rollout. This approach recognizes that AI, like any junior staffer, requires training, grounding, and a clearly defined scope of authority. By viewing AI agents as digital employees, Raleigh has successfully navigated the complexities of organizational culture, where fear of automation is often as significant as the technical challenges themselves.
Chronology of a Digital Evolution
The journey toward an AI-powered municipal help desk did not happen overnight. It was the result of a deliberate, multi-phase roadmap that prioritized data integrity before model deployment.
Phase 1: Foundations and the Rise of "Ral-E"
Before the introduction of sophisticated agents, Raleigh focused on the fundamentals. The city already possessed a robust data infrastructure and a long-standing commitment to the ServiceNow ecosystem. This foundation allowed for the creation of a "triage agent" affectionately dubbed "Ral-E."
Drawing inspiration from the animated robot of film fame, Ral-E was not designed for space exploration but for the mundane, high-volume task of sorting IT support tickets. Historically, this was a manual process that occupied significant staff hours. Ral-E automated the routing of these tickets, ensuring they reached the correct departments with the appropriate workflows and templates. With an accuracy rate of 98% to 99%, Ral-E proved that AI could handle routine administrative categorization with human-level precision.
Phase 2: From Triage to Action with "Alli"
Following the success of the triage agent, the city launched "Alli," a more advanced help desk agent. Unlike the strictly generative capabilities of its predecessor, Alli represents the shift toward "agentic AI." While generative AI creates content, agentic AI is designed to autonomously make decisions and perform tasks—such as adding users to distribution lists or executing complex password resets—with limited human oversight.
After a pilot phase that began in February, Alli is now fully available to city employees across all departments, operating 24/7 to provide instantaneous assistance.
Phase 3: The "Too Much Information" Lesson
The integration was not without its hurdles. During the training phase, Alli began recommending overly technical solutions to staff—such as suggesting manual registry edits that fell outside standard IT safety protocols. Wittenburg and his team realized that by granting the agent too much autonomy and access to a vast, uncurated knowledge base, they had created a digital employee that was technically correct but operationally dangerous.
"We realized we’d given her too much autonomy and too much information," Wittenburg admitted. The solution was to "ground" the model, retracting its permissions and slowly re-introducing data sets to ensure the agent’s advice remained within the bounds of city policy.
Supporting Data: Why Efficiency Matters
For government leaders, the primary metric for AI success is rarely "innovation for innovation’s sake." Instead, it is the tangible reduction of staff hours and the optimization of service delivery. Surveys indicate that 82% of state CIOs are currently exploring AI, with 40% of public sector leaders specifically measuring success through saved staff hours.
In Raleigh, the data backs the investment. Since the deployment of ServiceNow’s L1 AI Specialist:
- Service Desk Costs: Have seen a measurable decline.
- Staff Hours: A significant reduction in time spent on manual ticket triage.
- Routing Accuracy: The triage agent maintains a consistent 98–99% success rate.
- Future Goal: Wittenburg aims for AI agents to resolve up to 85% of all IT support requests, effectively liberating human staff to focus on high-value, complex problem-solving that requires human empathy and critical judgment.
Official Responses and Strategic Governance
The city’s success is not an isolated endeavor; it is supported by active participation in the GovAI Coalition. This group, which includes various public sector agencies, functions as a collaborative body for sharing AI innovations, navigating regulatory challenges, and setting best practices for ethical AI deployment.
Raleigh’s internal governance reflects this collaborative spirit. The city has developed dedicated AI training courses for staff, ensuring that employees understand what these models can and cannot do. Furthermore, a cross-functional group of department heads regularly evaluates the technology’s impact, ensuring that the deployment of AI aligns with the city’s broader mission.
Wittenburg emphasizes that the "cool factor" of a new technology is irrelevant if it fails to provide concrete value to the community. "I don’t want to pilot the coolest, latest thing," he noted. "Let’s pilot these things knowing that this is going to add value and is either going to make staff more efficient and effective or have a positive influence on the community."
The Implications: A New Era for Public Service
Raleigh’s experience provides a roadmap for other municipalities considering an AI transition. The implications of this model are profound:
- The Death of the "Manual Bottleneck": By automating the ticketing process, the city has proved that high-volume, repetitive tasks are the ideal candidates for AI. This shift allows human employees to pivot toward the community-facing initiatives that actually define public service.
- Data as the Primary Asset: The "learning curve" experienced with Alli served as a catalyst for cleaning up the city’s internal data. As Wittenburg noted, the mistakes made by the AI were often mirrors reflecting inconsistencies in the city’s own knowledge bases. In this sense, AI acts as a diagnostic tool for organizational health.
- The Rise of 24/7 Government: With agents like Alli, the concept of "business hours" for internal IT support is becoming obsolete. This sets a precedent for potential customer-facing applications, such as the proposed fully AI-powered customer kiosk at City Hall.
- Cultural Alignment: By framing AI training in the context of "onboarding a new employee," Raleigh has managed to reduce the anxiety typically associated with automation. This human-centric framing is crucial for long-term buy-in from the workforce.
Conclusion: The Path Forward
As Raleigh continues to scale its AI initiatives, the focus remains firmly on the community. The city’s journey underscores a vital lesson for the public sector: AI is not a replacement for human governance; it is a force multiplier.
"At the end of the day, we’re public servants, we are here to serve the community, and we’re here to make the community a better place," Wittenburg says.
For other cities, the takeaway is clear. The "AI journey" is not a sprint toward the latest headline-grabbing software; it is a methodical, iterative process of building a foundation, training the technology to reflect organizational standards, and ensuring that every automated ticket and resolved query contributes to a more efficient, responsive government. In Raleigh, the future of the public sector is already on the clock—and it is learning more every day.
