From governed AI to agentic workflows, starting with Fire & EMS scheduling.
.svg)
.png)
As "bring your own AI" spread across local government, the City of Hopkinsville, Kentucky, moved early to turn unmanaged, unsanctioned AI use into a governed program. With Darwin AI, the city gained visibility into how staff were using AI, stood up lean policy guardrails that protect citizen data without stifling innovation, and used Darwin Launchpad to build a fire-department scheduling workflow that cut a task once taking battalion chiefs three to four hours in a day down to about two minutes, with a human still in the loop.
Hopkinsville, Kentucky has become one of the most forward-leaning municipal adopters of AI in its region. Under Mayor James R. Knight Jr. and City Administrator Troy Body, the city chose to confront AI head-on rather than wait for a breach or fall back on a ban: it stood up an AI advisory board of department leaders, surveyed its own workforce, built a usage policy, and partnered with Darwin AI to put governance and real workflows in place across roughly 350 staff.
The city's fire department illustrates that complexity. The Hopkinsville Fire Department operates five Fire/EMS stations inside the city plus an EMS station in Oak Grove, with more than 100 firefighters and EMS personnel providing fire service for the city and emergency medical services for Christian County.
As AI adoption accelerated across local government, Hopkinsville recognized both the opportunity and the risk.
Hopkinsville didn't adopt AI in a single top-down moment. Like most agencies, its staff and the vendors working with the city were already bringing AI into daily operations to boost productivity and better serve residents. About a year and a half earlier, the city's Public Information Officer was using AI daily for communications work, and colleagues across departments were doing the same. The intent was good; what the city lacked was centralized governance. That gap created two potential areas of exposure.
The first was Shadow AI; the public, employee-adopted tools like ChatGPT and Copilot that staff reach for without oversight. As a local government, Hopkinsville holds information that citizens entrust to it: tax records, health-related details, and other personally identifiable information (PII). When that data flows into ungoverned tools, often by accident through an uploaded document or a copied email, the city has no way to see it or control where it goes.
The second was embedded AI; the AI features now baked into everyday software, from document assistants and browser extensions to the SaaS platforms staff already use, quietly processing city data in the background and often without anyone realizing AI is involved at all. In both cases the risk was the same: sensitive information moving through AI systems the city couldn't see, govern, or account for.
"There was real concern around AI use, especially around security and data. Darwin gave us a way to put guardrails in place while still enabling our teams to use AI productively." — Cody Brem, Policy Analyst, City of Hopkinsville
Rather than waiting for a compliance failure, the city established governance early. After City Administrator Troy Body attended an AI conference at Harvard, he returned convinced that AI was here to stay and that the city needed a deliberate strategy framed around opportunity: how can we use this to better serve residents, cut costs, and cut down the time it takes to deliver services?
He tasked the team with forming an AI advisory board composed of city department leaders. Its first realization was a humbling one: nobody actually knew who across the city was using AI, or how. So the board established a baseline, surveying staff about how they used AI and what concerned them. The results confirmed it: adoption was already widespread, across departments and use cases the board hadn't anticipated. That visibility turned AI from a blind spot into something the city could actually govern.
"We didn't want to take a reactive approach to AI. By establishing an advisory board and partnering with Darwin, we've been able to define how AI should be used across the city; securely and intentionally." — Amanda Brunt, Technology and Public Information Director, City of Hopkinsville
Hopkinsville saw two traps other cities had fallen into. Some wrote sprawling, restrictive policies that employees simply ignored, continuing to use AI on personal devices. Others stayed hands-off. Both pushed real usage into the shadows.
The city chose a third path: move fast, protect what matters, and empower employees to keep innovating. It deliberately avoided a 40-page rulebook nobody would follow, anchoring its policy instead in a few core principles; keep private information out of public AI tools, keep a human in the loop, and use responsible, sanctioned tools. Human review is the single biggest safeguard in the framework: before anything created with AI assistance is published, a person has to review it, driving compliance without heavy-handed restriction. With Darwin providing visibility into how AI is being used across the organization, that policy became something the city could actually monitor and enforce, not just publish.
"We wanted to make sure AI was being used responsibly across the city, not just efficiently. With Darwin, we're able to monitor usage, reduce risk, and give our teams structured workflows that actually improve how we operate." — Cody Brem, Policy Analyst, City of Hopkinsville
With governance established, the city turned to one of its most resource-intensive workflows: fire-department scheduling. Putting the right people, with the right certifications, on the right apparatus at the right station every day, while meeting minimum-staffing rules, was the hardest part of running operations.
Using Darwin Launchpad, the department built a scheduling model that plans a month out against the qualifiers that matter: a paramedic on every fire truck, swiftwater-certified personnel positioned ahead of incoming storms, and full accounting for vacation (booked up to 12 months in advance), sick calls, and the callbacks needed to hit minimum staffing. After staff submit availability, one button produces the schedule, and a human can still review and override it.
According to the city, the first phase has cut a task that took battalion chiefs roughly three to four hours of their day down to about two minutes, freeing up close to a quarter of their day. The department plans to expand the model from here. It's a clear illustration of the city's core lesson: value comes from grounding AI in a real business problem, not from deploying a frontier model and hoping magic happens. With Launchpad, Hopkinsville built that solution on top of its existing systems, without having to replace them.
"Darwin helps us ensure employees are using AI safely, while also giving us workflows that actually enhance day-to-day operations." — Payton Rogers, Fire Marshal and Public Information Officer, Hopkinsville Fire Department
Setting up the policy was only half the work; the harder half was getting people to use AI and to trust it. Fire and EMS are hands-on professions with real resistance to new technology, alongside broader anxieties about whether AI would replace jobs.
Hopkinsville's answer was to lead with problems, not pitches: rather than selling AI, the team showed each person how it removed a specific, repetitive task so they could spend that time on higher-value work; including AI drafting the narrative portions of run reports. Reframed as this saves you time, this speeds your response, skepticism turned into buy-in. On the formal side, the city partnered with a nonprofit public-sector AI program, free to the city, to push roughly two hours of ethical and generative-AI training out to all of its staff, covering what AI is, practical prompts, and the tools available to them.
"Our teams had legitimate concerns about AI, especially around cybersecurity. Partnering with Darwin allowed us to address those risks while still moving forward and using AI to save time and resources." — Amanda Brunt, Technology and Public Information Director, City of Hopkinsville
Governance at Hopkinsville extends to public accountability. For a Public Information Officer, transparency to residents is a core mandate, and AI has become a natural part of the city's story, briefed to the city administrator weekly and to city council quarterly, and communicated to the community through social media and other outlets.
The city has also thought carefully about how AI intersects public-records law, an emerging gray area nationally as media outlets file requests about which AI tools agencies use and how much AI-assisted drafting must be disclosed. Working with legal counsel, Hopkinsville found that under current Kentucky case law, working drafts and search-style interactions are generally treated more like Google searches than disclosable final records. The city expects this area to grow more complex and anticipates closer scrutiny, and audits, as AI's role expands.
Hopkinsville demonstrated that cities don't have to choose between innovation and control. With the right governance layer, early visibility, a lean enabling policy, and workflows grounded in real problems, AI can be secured, trusted, and put to work in frontline operations. Asked what they'd do differently, the team's answer was unanimous: start sooner. Risk mitigation will always be there; the way through it is to begin the conversation, survey your staff, and start building.
Schedule a demo to discover how quickly you can gain control of the AI in your agency.
.png)
