AI that delivers for your city

Bill Eggers: From Deloitte, this is Government’s Future Frontiers, the podcast that asks questions today to help create tomorrow.

I’m Bill Eggers, the executive director of Deloitte’s Center for Government Insights, and in this episode, you’re going to hear a conversation recorded live in Barcelona, Spain, at the Smart City Expo World Congress.

It’s an influential event on urban innovation where key players from all over the world come together. And that makes it an ideal place to explore this episode’s subject: AI-powered cities and leading practices in enhancing performance at city hall.

You’re going to hear my conversation with two innovators: Rochelle Haynes from What Works Cities and Suma Nallapati from the city and county of Denver.

You’ll learn about some of the latest real examples of how AI is transforming life for city residents, and my guests explain how they see artificial and human intelligence working together to create a better future for all.

Eggers: Well, welcome, we are at Smart City Expo World right here, with thousands and thousands of smart city enthusiasts. And I am really excited about this panel today because it’s with two individuals, innovators I’ve admired for a very long time. So, the panel we are going to have you two talk a bit about yourselves. But first, we have Rochelle Haynes from Bloomberg Philanthropy, What Works Cities, Results for America, and then we have my friend Suma Nallapati who is now the chief of AI and information officer for the city of Denver. And why don’t we start off by you two saying a little bit about yourselves and what your roles are, and so on?

Rochelle Haynes: Wonderful. So, Bill, first and foremost, thank you so much for having me. I’m the managing director of What Works Cities. What Works Cities is the international standard of excellence on what it means to be good, well-managed local government. I’m excited to be leading this work. I’m a former public servant of New York City, where I focused on affordable housing development, social services, and homeless services work. And so this role allows me to combine my experience in both the public sector work as well as the nonprofit and philanthropic sector.

And what’s exciting about What Work Cities and what we’re doing right now is that we are connecting city leaders with the information, the data, and the technical assistance skills they need to be more data-driven and evidence-based in their approach, but also be ready for the future—and, which we all know, AI is a big part of that future.

Eggers: Thank you. Suma?

Suma Nallapati: Yep. Speaking of AI, it’s everywhere. And I’m very proud to lead the city and county of Denver under Mayor Mike Johnston—the AI strategy, along with our technology strategy for the city and county of Denver. I’ve been public sector and private sector. I’ve worked with Deloitte in my public sector roles and also in the private sector, and also Bloomberg Philanthropies. We just concluded our Denver AI Summit focused on public sector challenges and where AI can play a part. So very excited to be part of this amazing conference with so much thought leadership. And thank you Bill so much for having us.

Eggers: How did that come about, your new title of chief AI officer?

Nallapati: It’s very symbolic from the work that’s being done. Our resources are constrained, right? The budgets are constrained, so AI is a tool in our toolbox to help us leverage capabilities, which may not be offered to us otherwise. So, very excited. And also responsible AI is important. Data security, privacy—all those are very critical in this framework. So, the title is a culmination of a long, hard-fought journey, being in IT for almost 30 years. I started out with machine learning, deep learning, large language models, and this is a natural progression. It’s again: How do you take data and use the outcomes to help with residents and their resident experience? So it’s all about the experience for our residents.

Eggers: Wonderful, Rochelle. What Works Cities has been around government performance at the city level, getting certified, looking at how to improve that performance. And now you’re looking at expansion. You’ve expanded into South America, a little bit into Canada. Talk a little bit more about What Works Cities, what it does, how do you get certification, and what you’re looking at doing from an expansion perspective?

Haynes: Absolutely, and first and foremost, I want to acknowledge Bill. Bill is one of our What Works Cities Standard Committee members, and he’s been a champion of our work from the very beginning. And as I mentioned, What Works Cities is an international standard of excellence. It really helps cities benchmark their assets as well as their risks.

We’re taking a look at things like: How do you manage your data? What does your data governance policy look like? Are you building evaluations into your processes and your policies that you’re designing? And for cities, we’re not just asking you to map these assets and risks. We’re combining that with free technical assistance that we provide from experts, partners throughout the United States, as well as Latin America and Canada, to really help you build that internal capacity.

We currently have over 100 cities that have certified with our program in North, Central, and South America. We’re thrilled about that. And what’s really exciting, I have over 220 cities in the network, over 1,800 city leaders, so this is a global network of leaders around the world that are connecting with one another on how to drive data-driven decision making, how to make decisions that connect to the residents, and also how to get at the root of the challenges that you all are facing. And what’s powerful about the network is it’s global. And what you find is the conversations you’re having in the United States aren’t that different from Latin America. We’ve created a space for leaders to connect with one another. So, it’s a network of doers, as I like to say, we’re networkers of doers and collaborators that are sharing best practices with one another.

Eggers: Fantastic. And I’d love to ask this question to both of you. What’s on everyone’s mind right now, of course, is AI. So smart cities and AI are, basically, you know, very, very connected. Now, Suma, you have an early background in that, and I remember I first started writing about kind of AI in government in a book over 20 years ago, and those were rules-based engines and not nearly as sophisticated as what we see today. Can you talk a little bit about the role that AI is going to play in government performance and innovation, and helping citizens?

Nallapati: Absolutely. Housing, affordability, government inefficiencies. Right? Why does licensing and permitting take as long as it does? Can AI be applied, right? There’s all these complex rules around all of that, that human intelligence needs to be augmented by AI in this scenario. So I’m very excited. We’ve started actually with the information governance. That’s important. Policy is important. Responsible AI is important, but once you establish those guardrails, the potential for AI is limitless in my mind.

How do you take the transformational work and give it to humans? They want meaningful work. Our employees want meaningful work. They just don’t want to do the repetitive, mundane tasks. They want to use their intelligence in ways that will help our residents. So do that with human intelligence, but leave the transactional work to the bots.

So that’s where the repetitive mundane tasks get eliminated in our workflows, and we focus on the truly meaningful work. We’ve been able to do quite a bit already. We’ve introduced this platform, which sits on our Denvergov.org website. It’s called Sunny. It’s integrated into all our backend system using data. Our teams are curating it for hallucinations. It doesn’t read off of the internet, right, as an example. And we just look at ways to help our residents. Sunny answers questions in 72 different languages. And Bill, if I have to hire 72 different translators, it’s simply impossible. Right? Like people that may be hungry, right? They can give up a lot of things, but if they have their phone, what we are seeing is, “I’m hungry. Is there like a place where I can get food tonight?” And they’re asking that of Sunny. For me, that’s powerful because they may feel a lot of shame in calling a call center, but they’re able to do that much more easily on the platform, and it creates a case. We follow-up immediately, and it’s a much more compelling platform than just someone calling on the phone. So we are finding ways, again. We worked with an MIT startup to have a licensing and permitting software integrated into our whole process, and it’s getting us more than 30% efficiencies already.

Eggers: That’s wonderful. I think a lot of the regulatory areas and permitting and even housing and using digital twins and what’s called regulation as code, where you’re bringing that in and even around procurement. There’s so many ways to both speed things up for citizens and businesses, but also to do what we call scaling the human edge. Really focusing on what humans can do better, and so on, and taking away some of the manual tasks.

Now, Rochelle, you spoke on the main stage yesterday on AI and AI-powered cities. Let’s hear a little bit of your thoughts on this, and especially as it connects to what you’re doing at What Work Cities.

Haynes: So yes, I had the pleasure yesterday of participating on the AI and Urban Transformation conversation. What I love about that conversation and all the conversations we’re having about AI is that we’re centering cities. Cities are at the forefront of the global revolution, and they’re going to be at the forefront of the AI revolution.

Cities have the opportunity to do a really big thing right now and revolutionize how they show up by leveraging AI technology. And so for me, what I’m excited about is that cities are actually trying AI. I think, in the past, sometimes public sector’s a little bit reluctant to adopt technology, but there is a willingness amongst cities and leadership within cities to try.

What I find is that cities are most comfortable doing pilots that feel really practical right now. So how do you practically help me process permits and applications? How do you practically help me do preventative work, like detect wildfires? We’re seeing that in Austin, Texas. They’re using AI technology to detect wildfires. In Recife, Brazil, they’re using AI technology for flood prevention. But what’s great about all of these is they’re connecting with their residents. They’re not building it in a vacuum and making citizens think that this is some sort of big brother thing that’s behind the scenes. They’re being upfront and transparent and cocreating with residents.

And so for me, when I think about this [00:11:00] moment in cities, I think this is a moment to revolutionize how cities operate. I think it is going to free up time, as you mentioned, for city staff, and it’ll free their time so they can be creative and innovative, and not just in crisis mode.

I’m a former public servant. We spent more time sometimes in crisis mode than we did in innovation. I think AI is going to create the space for cities to have that staff, to have that space to be more innovative and creative with how they think. And then I think on the other side, it’s an opportunity to democratize information for residents. So residents can access information within their city halls. They can get the answers to the questions if you need the meal for the night, but it also allows government to be more transparent. And when you have a transparent government, you can have a citizenry that is actually engaged.

And so I’m, as you can tell, very excited about what AI can do. And I think there’s work to be done, right? To clean up data, organize data, make sure we’re addressing biases, but at the same time, we don’t have to let the perfect be the enemy for the good. We can try and test in a way that feels safe and comfortable and engages our residents.

Eggers: Well, let’s get into the data piece a little bit, because that’s a big piece of the AI puzzle right now. And we’ve heard from a number of folks here over the last two days just about big problem with the data silos, both within different government departments, levels of government, but also between the public and private sector, where at a city level, a lot of transportation data—a lot of other data—sometimes is held by the private sector. What do we do about the data piece of this and being able to start to break some of those silos so we can get the most out of AI? And Suma, you’ve been in this area a long time, so I’d love to hear from you.

Nallapati: I really like what you [Hayes] said, right? Let’s not wait for data to be perfect before you start the AI journey. I think they both can coexist. We can learn from each other on those two tracks. Like, how do you start the journey? How do you make your AI and iterate? That’s important.

Bill, to answer your question directly, let’s start from the resident and work our way backward, rather than having to try to merge all these data silos and stuff. Let’s see what the resident wants. Ask the questions and then build your data pipelines to match that. And you want to meet the residents where they are.

And again, when they’re interacting in the private sector, they’re interacting with very sophisticated AI. It’s curating the content based on their preferences. And when it comes to public sector: Why is it so difficult, right? That’s the question to ask.

What are those interactions that are most common? Start with that. Start with an MVP, start with the pilot, of those data attributes, and then go from there. So that’s the only way to get started. And, Bill, as you know, I worked in the state and we built “My Colorado.”

And again, with My Colorado, the app was the easy part. How we got the data pipelines behind was the most time-consuming. But learning from those kinds of experiences, it’s all about the resident. And then that makes the equation easier.

Eggers: That’s a great explanation. And Rochelle, what are your thoughts on the data piece?

Haynes: So one of the things you touched on was residents, right? And so we know that sometimes data is flawed or can be biased. Old data sets. This is a moment to clean that up. Use the data that you have, but at the same time, find accessible new collection methods that allow residents to engage with you and collect new data. And one example I’ll highlight this, like in action, in a city that’s part of our network—that’s Fort Lauderdale, Florida.

They were going to do investments in stormwater. There was excessive flooding. Before they did the investments, they did GIS mapping to identify where there are floods. But then someone said, “Hey, let’s also have focus groups and conversations with residents so residents can tell us where there’s flooding.”

And to no one’s surprise, beyond just the GIS data, they identified I think it was like 25 additional sites, where there was flooding that data alone didn’t pick up. So it’s this combination of the resident engagement data as well as the hardcore data that you can collect through methods that allows you to be more informed.

And I think, and that’s the approach we need to have with AI. It’s like, let’s use what we have get started, but use this as a moment to clean that data up. But I think, at the end of the day, we don’t have to wait for it to be perfect. And I think that’s the biggest takeaway. I think some of the messaging before has been everything has to be perfect before you start.

It’ll never be perfect, but what you can do is start to leverage small pilots as the conversations that you can have internally to do this. And that’s really also how our certification is designed. You can’t get certified without talking to one another. It’s purposely designed that way, so it creates that interconnectedness.

Eggers: That’s wonderful. Suma, you just launched Denver’s Sunny chatbot. And I assume it’s called Sunny because Denver has more sunny days than any city. How is that going, and what does it do?

Nallapati: It’s a great platform. The beauty of Sunny is how well it’s integrated to our backend systems, right? So it’s not just a chatbot—[that is,] front-facing—it’s integrated to the backend. Technology is working behind the scenes. We have had more than 40% of our call volume for 311 go through Sunny now. That’s within one year.

And again, when I say resources are constrained, that 311 team hasn’t grown. But our interactions have gone up, and Sunny’s able to augment the staff, right? And it’s going extremely well. CIO 100 award-winning platform. And a lot of other cities are coming to us on how we got started with that.

And it’s been truly one of the biggest things for us in the recent past with how we were able to integrate the data behind the scenes, clean up the data, but also make it very easy for the residents to interact with city government.

Eggers: Well, congratulations on that because a lot of the early uses of AI have been sort of back office and everything, and people were afraid to do things citizen facing too early.

Well, we’re almost done, but I want to ask one last question to both of you. Looking out 10 years from now, how do you think cities will be transformed by AI, by digital twins, IoT, even quantum computing, and other emerging technologies?

Haynes: Oh, the cities of 10 years from now, I will say they will be labs of innovation. I think AI is going to automate tasks that should be automated. AI is going to free up staff time to have the space to think big—about policies, about approaches and frameworks, and innovative design. I think AI is going to have city halls where citizens are able to get the resources they need in a more efficient way, and I think it’s going to be an exciting time, for all of us.

I think that, you know, in some ways I am at the five-year mark, not even a 10-year mark, because I think there’s probably not even the evolution of the AI technology on how [00:19:00] far it can take us with city halls, but I think it’s going to be a city hall that is deeply responsive. Think about a city hall that has a dynamic dashboard that can predict flooding that can get ahead of the wildfire, that can use a digital twin to build affordable housing in a different way and like be able to use that digital twin to explain it to their residents.

Eggers: And to explain the trade-offs

Haynes: Exactly. And explain the trade-offs and so citizens feel more engaged. And I think that’s where we’re headed, and so it’s really exciting time.

Eggers: Great. Suma?

Nallapati: I would say, Bill, the recursions are happening so fast with AI. I can’t predict how it’s going to be for the next six months, right? It’s happening so very fast, which is all great. How do we keep up with that pace? We all have to increase our literacy, our awareness, our education, and all of that. I think with all of it, my hope is that we serve more residents. Maybe our health outcomes are better. Maybe we get better water, better living, better resources for our residents where they truly feel that they belong in the city and that they are being serviced with the right resources, whether it’s AI or something else.

We democratize the data to more effectively service our residents that demand more. They want more. They deserve more. And I think AI is going to get us there.

Eggers: Well, I’m also looking forward to flying taxis because I hate being stuck in congestion, and that’s been on the radar for 10 years. I guess you can go to Dubai.

Haynes: I support that, as a native of New York City and who sat in traffic many meeting days. I support that.

Eggers: Well, thank you two very much. It’s so good to see you and have a great rest of the conference.

Haynes: Thank you.

Nallapati: Thank you.

Eggers: Well, that’s it for this episode. We are out of time. Thanks to Suma and Rochelle for such an inspiring discussion here in Barcelona. If you’d like more from Government’s Future Frontiers, you’ll find all our previous episodes wherever you get your podcasts. And to make sure you don’t miss new episodes, be sure to follow the show on your favorite podcast platform.

This podcast is produced by Deloitte. The views and opinions expressed by podcast speakers and guests are solely their own and do not reflect the opinions of Deloitte. This podcast provides general information only and is not intended to constitute advice or services of any kind. For additional information about Deloitte, go to Deloitte.com/about.