The most meaningful AI wins often start with the most ordinary work.Suma Nallapati, CIO of the City and County of Denver, believes in starting with solving concrete problems within the organization — not just instituting “AI for AI’s sake.”
“The power of technology,” she told me in a recent AI-First Podcast episode, “is helping the most vulnerable populations in the most vulnerable times. And if a technology can help even one human being, my job is done.”
The organization, which operates 53 different agencies, recently introduced Sunny, a secure, encapsulated virtual assistant built directly into the city’s website. Named after Colorado’s famous 360 days of sunshine, Sunny is AI-trained on Denver’s internal knowledge bases and integrated with backend systems like Salesforce using MuleSoft.
Sunny allows residents to instantly find answers to their questions 24/7 and generate service cases in 72 different languages without needing to wait for a human agent. For vulnerable populations — like those seeking immediate shelter to avoid homelessness — Sunny provides a private, dignified, frictionless way to access critical services instantly, transforming how the city interacts with its residents.
Here’s a longer look at how AI is helping Denver reinvent its customer service, and the lessons public-sector workers everywhere can take from their experience.
Key takeaways:
- There’s an emerging best practice to getting started with AI: Begin with trusted content, clear guardrails, and a problem employees actually need solved
- Enterprise AI needs governance and context: AI works best when content is secure, organized, and trusted
- Early AI value comes from reducing routine work: Helping with repetitive tasks like policy questions and document review
- AI should support experts, not replace them: Its ideal role is to surface relevant information faster so people can apply judgment
Accelerating services while protecting sensitive data
We all know AI can move faster, but can it do so without breaking the rules that matter most?
In the public sector, the stakes are incredibly high. Deploying AI in municipal government requires absolute trust and data security. The tension of capitalizing on AI without exposing content to risk was one of the first things I asked Nallapati about.
We need to know that we’re safeguarding our residents’ data with utmost ethical views, so we have a very robust information governance committee that looks at policy, ethics, transparency, and explainability.
She emphasizes the importance of maintaining strict ethical standards when handling public data: “We need to know that we’re safeguarding our residents’ data with utmost ethical views, so we have a very robust information governance committee that looks at policy, ethics, transparency, and explainability.”
Supporting these efforts, Box provides a secure, centralized content foundation. Enforcing strict cybersecurity standards and using the automated controls Box provides helps the city catch and block PII before it can enter public training models. This secure environment allows IT leaders to be fearless in their experimentation while keeping resident data protected.
Speed matters. But trust matters more.
AI doesn’t replace Denver public servants — it augments them
Nallapati describes the AI-powered workflow design Box enables as a collaborative feedback loop that uses technology to handle transactions and humans to handle transformation. This construct reinforces that the best AI systems don’t eliminate human judgment but instead improve the path to it.
AI can accelerate orientation, reduce prep time, help people speak the language of a domain faster. But the final standard still belongs to humans, and offloading routine administrative tasks lets city employees focus on the core mission that drew them to public service in the first place.
We’re able to get the transformation work to humans and leave the transaction work to bots.
“A lot of people join the public sector because they want to give back,” Nalapati notes. “When you take the repetitive, mundane tasks out of their workflows, it drives them to be more innovative, and that creates momentum as well. We’re able to get the transformation work to humans and leave the transaction work to bots.”
By automating routine transactions, the city has achieved major efficiency gains. In traditional workflows, a human-handled 311 transaction costs the city $4.75. After Denver’s move to the Sunny system, the cost has plummeted to just $0.35 per transaction. These savings allow Denver to maintain strict service level agreements (SLAs) despite a 30% budget-driven reduction in customer service staff.
Where AI delivers value first: Repetitive, time-consuming work
Thoughtful AI adoption should be practical, grounded, and relentlessly focused on helping public servants spend more time on the meaningful work they were hired for.
Some of Denver’s most valuable AI use cases, Nallapati insists, aren’t glamorous at all, but operational and procedural. They’re the kinds of tasks people quietly dread, like navigating complex business licensing and permitting. City agencies generate massive volumes of structured and unstructured data. Armed with AI, employees who used to spend hours searching through legacy systems can now instantly retrieve policy guidelines and resident records.
The role of AI isn’t to replace the reviewer but to amplify the signal, reduce the noise, and get the right human to the right passage faster.
The same pattern shows up in document review. Denver’s Civic Check workflow, for instance, which was curated directly from the city’s annual AI Summit, helps businesses prepare paperwork and navigate complex permitting rules before they enter the official system, cutting down on back-and-forth delays.
The role of AI isn’t to replace the reviewer but to amplify the signal, reduce the noise, and get the right human to the right passage faster. If the task is repetitive, rules-based, and time-consuming, that might be exactly where AI can help first.
Content curation makes for better AI input
Nallapati offers another important perspective on AI: You can’t throw a thousand unrelated documents into one big pile and expect AI to surface excellent answers. If the data spans different time periods, authors, and subject areas without clear structure, the outputs will get messy fast.
She argues instead for curation: Better metadata. Better descriptions. Better rules. Better expectations for output. The quality of the AI experience depends heavily on the quality of the information environment around it. Metadata, categorization, and context aren’t back-office chores. They’re what make intelligent systems reliable.
This emphasis on curated, organized content is right in line with the Box ethos on Intelligent Content Management, which I see confirmed by every customer we serve. Those who take the time to organize and govern their content are naturally set up to benefit from AI.
What should leaders take away from this?
Used well, AI is leverage that helps people get oriented faster, find the right sources sooner, and spend more time on work that actually requires judgment, creativity, and care. The pilots Nallapati values most are the ones that help save time and resources while protecting vulnerable populations.
I think that’s a better north star than most grand transformation slogans. When you remove drudgery, you create room for better work. And when you pair AI with curation, governance, and human review, you create the conditions for trust.
The future of AI at work may look less like a dramatic handoff from humans to machines than like a disciplined redesign of how knowledge gets found, reviewed, and used.
For more from Nallapati, I invite you to watch the entire episode of AI-First.