Spotlight on Service: Building the Future of AI for Government Today with Santi Garces

Celebrating and learning from those on the frontlines of public service is core to our mission at Civic Roundtable. This month, we sat down with Santiago 'Santi' Garces, the City of Boston’s Chief Information Officer. 

‍Thanks for speaking with us, Santi! I’m excited to learn more about your background, and how you got into the work you do.

Santi: As people might be able to tell by my shirt (a Colombian National Soccer jersey), I was born in Colombia and came to the U.S. nearly 20 years ago. I studied electrical engineering and political science before earning a master’s in Entrepreneurship in Science and Technology from the University of Notre Dame.

After trying to build a biomedical startup, I found myself inspired by a new mayor in South Bend at the time, Pete Buttigieg. I had the opportunity to help build the city’s innovation and performance program, later worked in Pittsburgh, then returned to South Bend as Executive Director of Community Investment, overseeing housing, zoning, and workforce development. Four years ago, Mayor Michelle Wu asked me to join the City of Boston as CIO, and I’ve been here ever since.

I have lots to ask, but maybe I’ll start with some fast facts — what's your favorite TV show? And any pump-up songs on your radar for the World Cup?

Santi: I've been rewatching Silicon Valley, which has been really funny. And I've been listening to the new Shakira song — that's the World Cup official song. And then, obviously, there's a Colombian hype song for the Colombian national team that I play on repeat.

Now, jumping in. Where is AI creating the most tangible value in government today?

Santi: So much of our work in government involves consuming large volumes of information – complex regulations, policies, and guidance – and then producing documents. A bureaucracy is a perfect place to have an LLM. 

Some of the work that sounds a little tedious – like summarizing policies, analyzing regulations, drafting memos, helping interpret data – but it is actually where these tools shine.

Government is constantly being asked to do more with limited resources. If AI can help public servants spend less time on repetitive work, it creates more capacity to respond, adapt, learn, and improve. That’s my optimistic view.  

If we get this right, what does government look like in 10 years?

Santi: Governing is ultimately a shared learning exercise. The beauty of a democracy is we get to call the shots, we get to make decisions. At its best, government will spend more time making the complex decisions that require human judgment. AI can help us talk with each other more effectively, compress and expand information in ways that are very efficient. 

As governments invest in AI, where do you see the biggest mistakes? And where do you see the biggest opportunities?

Santi: I think we're in this messy period where people are very fixated on use cases. That’s understandable since you want projects that demonstrate measurable value. 

We should also be thinking about platforms. How do these things scale? Cities and states are very heterogeneous organizations. The same AI tool won’t work equally well for a field inspector, a policy analyst, and a data scientist. Different people do very different work, and it's unlikely that the same tool is going to be useful for everybody.

The bigger opportunity is building flexible platforms that can support all of those different ways of working instead of optimizing around a handful of isolated use cases. Long term, I expect LLMs themselves to become commodities. The real value will come from how organizations build sustainable platforms around them.  

Collaboration is one of the hardest parts of government. How do you approach it as Boston’s CIO?

Santi: I think there's nothing worth doing that doesn't require some level of collaboration. Especially in cities, the issues we experience are multi-dimensional. 

For example, crime is a public safety issue, a public health issue, and an issue of street design.  Do we have dark corners that are easy for people to hide in, do we have good, clear, visible spaces? It’s very multidisciplinary by nature, and strong collaboration is important.

Tools are really critical. If it's hard for people to work together across different teams, that creates barriers. Culture matters. Institutional processes matter. Sometimes there are legal barriers that prevent information sharing. People, process, and technology all have to come together if you want effective collaboration.

When an effort or initiative touches multiple departments today — like a 311 request that touches public works — what does it look like today from an IT and tech perspective?

Santi: We think about city work through the lens of service management and project management. We’ve built an inventory of city services, and many of them are actually fairly linear. A pothole is a good example. Is it our road? Do we know where it is? If yes, patch it.

The piece that becomes difficult is when you're working with people across organizations. By nature, our department works with almost every other department because we’re providing technology, solving problems, and trying to bridge the seams that naturally exist inside a large bureaucracy. 

Cross-department efforts are dynamic. You’re constantly bringing together different people and resources to solve evolving problems. Those pieces are tough, really complicated, not linear,  they're about managing and bringing resources dynamically to try to address an issue. We're excited to have great entrepreneurs like Civic Roundtable building tools that allow you to address this. 

We hear a lot that AI is only as good as the context it has. How do you approach creating sources of truth for data in the city?

Santi: The first step is recognizing that this isn’t a one-time project, it’s about continuously building organizational maturity. 

One example is permitting. We realized we simply didn’t have enough content explaining how residents complete common projects like installing solar panels or renovating their homes. If people can’t find clear information themselves, AI certainly won’t be able to produce reliable answers. 

Recognizing that, we started by building better content. We used AI to help create an ontology of the permitting experience, a structured map of roughly 260 different resident experiences that relate to permitting. Once you have that foundation, AI becomes incredibly useful for helping people maintain content, organize information, and keep everything current.

The last piece is retrieval. Technologies like vector search and embeddings make it much easier for people to find semantically relevant information instead of relying on exact keyword matches.

We think about AI across four areas, strategy, content creation, content management, and retrieval. It has the potential to improve every part of that process.

Are there any projects you’re especially excited about right now?

Santi: We’ve been doing some exciting work around our open data portal by making it available through an MCP server. It opens the door for both city staff and the public to interact with city data in entirely new ways. 

We’re also building a digital public infrastructure stack around constituent identity. We’ve already launched the first version through our vendor registration system, where people can create a reusable digital identity that will eventually work across multiple city services.

As AI agents become more common, secure identity becomes increasingly important. If someone wants ChatGPT to pay their parking ticket one day, we need secure systems that verify they’ve actually authorized that action. Building those security foundations now is incredibly important.

Boston is an amazing place for innovation, and we're lucky to have such a great ecosystem. We're really proud and lucky that you guys are here in Boston, that Civic Roundtable started here in Boston. We hope you continue to build the momentum, establishing Boston as the place for good, substantial technology that's changing the world. 

Thank you, Santi, for taking the time to share your perspective. We appreciate your leadership, and everything your team is doing to help shape the future of AI in government.