Oh, hi!
Denver just pulled the plug on its AI chatbot.
The Denver City Council rejected a nearly $4 million contract extension for Sunny, the city’s AI-powered chatbot. Council members raised concerns about accuracy, its integration with 311, and whether the tool was actually delivering the experience residents needed.
Denver was trying to solve a problem every local government knows well: How do we make it easier for residents to get information and services?
That is exactly the kind of problem AI should be able to help solve.
But Denver’s experience is also an important reminder that implementing new technology and improving the resident experience are not the same thing.
The chatbot isn’t the outcome
When I oversaw digital services and 311 in Gilbert, Arizona, one of the things I learned quickly was that residents don’t care which department owns their problem.
They don’t want to know whether a pothole belongs to Public Works, whether a missed trash pickup belongs to Environmental Services, or which department maintains the park down the street.
They just want it fixed. They want their services improved.
That sounds obvious. But government technology has historically been designed around the organization rather than the resident and thinking about them first.
And AI gives cities an opportunity to change that.
A resident should eventually be able to say, “The streetlight outside my house has been out for three nights,” and government should be able to figure out what happens next.
The resident shouldn’t have to find the right webpage, identify the right department, choose the correct category from a dropdown menu, or call three different phone numbers. Who wants to make a phone call these days?
That is the promise of AI in resident experience.
But getting there requires us to measure the right things. Stop measuring activity. Start measuring outcomes.
It would be easy to measure an AI chatbot by how many conversations it handles.
10,000 questions answered. 50,000 interactions. 20 percent fewer calls to 311.
Those numbers might be useful, but they don’t necessarily tell us whether the experience got better.
Cities should also be asking:
Did the resident get the right answer?
Did their issue get resolved?
How many times did they have to contact us?
Did AI reduce the amount of effort required from the resident?
Did it make government easier to navigate?
And maybe most importantly: Would the resident choose to use it again?
Those are resident experience metrics.
There is enormous pressure on cities to adopt AI right now.
There will be pilots. Chatbots. Copilots. Agents. New platforms promising to transform government services.
Some of them will be extraordinary. Some of them won’t work.
That’s okay.
The mistake would be defining success by whether a city implemented AI at all.
The better question is much simpler:
If the answer is yes, keep going.
If the answer is no, the technology isn’t the innovation. It’s just another tool residents have to navigate.
Talk soon,
Dana
About Dana
Former Emmy-winning television producer and Chief Digital Officer. Built Gilbert, AZ's national award-wining Office of Digital Government. Now helping city communicators tell stories that build trust.
Oh, hi! Stories Podcast: Spotify | Apple Podcasts
Trusted by city leaders, PIOs, and civic innovators across the country.
{{rp_refer_url}}