Today, that front door is expanding beyond pages and menus and into conversational, always-on digital experiences. During a recent webinar with the National Association of Government Web Professionals (NAGW), Granicus leaders and local government practitioners explored how CMS platforms and digital agents are reshaping citizen engagement.
What emerged was a clear picture of where government digital strategy is headed, what challenges remain, and why web professionals are uniquely positioned to lead this next phase of transformation.
From “trendy” chatbots to mission-critical digital agents
Not long ago, chatbots were viewed as experimental. They were interesting, but difficult to scale and maintain, said Luke Norris, vice president of platform strategy and digital transformation at Granicus.
That perception has changed.
“With the emergence of generative AI, we’re certainly seeing that conversational digital agents are absolutely trending,” Norris said. Governments today are “being stretched and asked to do more with less,” while residents increasingly expect conversational, self-service digital experiences that mirror those in the private sector.
Norris noted that residents picking up the phone to call government offices are often seeking basic information that is too hard to find online or not properly personalized to address their concerns. In fact, “60% of those call volumes really are relatively simple questions,” he said. Digital agents now offer a way to meet that demand while improving efficiency and access.
Trust, accessibility, and the government advantage
One of the most striking insights shared during the discussion was how residents perceive government-provided digital agents compared to commercial tools.
According to Norris, Granicus research found residents were “more than twice as likely to trust conversational digital agents that are actually provided by the government compared to commercial LLMs.”
Trust matters most for high-stakes interactions “like benefit programs, tax, things that residents have to do to comply with government services,” he said.
Accessibility also emerged as a top priority for government web teams. Polling during the session confirmed what many agencies already feel: accessibility requirements are accelerating, often without additional funding. Digital agents can help bridge that gap. Norris pointed out that conversational interfaces allow residents to “ask in plain language” and receive responses “at a 5th or 6th grade reading level,” reducing barriers for many users.
These agents can also mitigate challenges related to screen readers and navigation by summarizing content and reducing reliance on complex page structures, making them a complementary layer to traditional accessibility efforts.
Moving beyond proof of concept to proof of value
While interest in AI is growing, both speakers cautioned against rushing implementations without clear goals. Marcus Townsend, east regional director for NAGW and solutions architect for the City of Fayetteville, N.C., described how early chatbot deployments sometimes fell short, noting they were “not overly excited about it” and concerned about long-term value.
Norris emphasized that many early failures stemmed from focusing on “proof of concept” instead of “proof of value.” Successful deployments, he said, are grounded in measurable outcomes such as reducing call volume, improving accuracy, or increasing access to personalized information.
Another risk factor is attempting to build and maintain AI systems internally without sufficient resources. Norris said Granicus employs “27-plus data scientists” to continuously test, tune, and optimize models, an investment that most local governments cannot replicate. As a result, many agencies are moving away from “homegrown or stitched together capabilities” toward vendor-hosted solutions explicitly designed for public-sector use.
Why data and content governance still matter
Despite the power of generative AI, the conversation made one thing clear: digital agents are only as effective as the data behind them. “The LLMs are not magical,” Norris said. “It’s all about data at the end of the day.”
Too often, leaders think of data only as spreadsheets and databases, overlooking the critical role of website content. “They don’t understand or think about data as the actual data on your website,” Norris explained. Outdated pages and poorly governed content can all undermine digital agent performance.
At the same time, digital agents can expose hidden issues. Some governments discovered inaccuracies only after feeding call center scripts into their agents and realizing those scripts were outdated. In that sense, conversational AI serves as both a service and a diagnostic tool, helping agencies improve the quality of their underlying content.
The future of the government website experience
One of the most forward-looking parts of the discussion centered on how digital agents can reshape the very structure of government websites. Townsend posed a question many web professionals are quietly asking: “Are websites even going to look like they do in the near future?”
Norris believes CMS platforms will remain essential for creating and structuring content, but the user interface will evolve. He described a future where residents interact through “a conversational experience” layered over traditional pages, one that users can exit at any time to access deeper information.
For the next several years, Norris expects residents to move fluidly between web and conversational experiences. “Sometimes they’ll start conversationally, but want to move to a web[page], and then come back,” he said. That hybrid model reinforces the importance of user experience design and customer journey mapping — skills that web professionals already bring to the table.
An opportunity for web professionals to lead
As governments navigate AI policy questions, funding constraints, and organizational silos, one message stood out: web professionals are well-positioned to lead this transition. Digital agents rely on the same foundations that web teams already steward — content quality, information architecture, accessibility, and user experience.
Norris encouraged agencies to reframe AI not as a risk, but as “augmented intelligence,” an extension of the automation governments has used for years to improve access to information. “It’s not going anywhere,” he said, underscoring the importance of getting started now and shaping how these tools evolve.
In the end, leveraging CMS platforms and digital agents isn’t about replacing websites or staff. It’s about creating more responsive, empathetic, and accessible government experiences that meet residents where they are, whenever they need help.
Want to see how these lessons play out in the real world? Read how Jackson County, Missouri, is using AI-powered digital agents to deliver a better public service experience for its community.
