Governing Smarter: How AI is Advancing Modern Cities

EVOLVE 2026 · MIAMI · PANEL · SESSION 02

City and county leaders from New York, Baltimore and Miami-Dade discuss what it takes to move government AI from pilot to practice—and why leadership, governance, and trust matter more than technology alone.

Evolve is Sand’s annual summit bringing together technology leaders, policymakers, and practitioners working at the intersection of AI and critical infrastructure.

Session 02 from our 2026 Miami edition tackled one of the hardest problems in public-sector AI: why successful pilots so often fail to become permanent systems. While the conversation covered everything from data sharing to procurement, one message emerged above all others: deploying AI is no longer primarily a technology challenge—it’s an institutional one.

Watch The Full Session

THE 60-SECOND READ

Governments aren’t short on AI ambition or on data. They’re short on the governance needed to deploy it at scale.

Bringing together leaders from New York, Baltimore and Miami-Dade, this discussion explored what separates successful AI initiatives from the countless pilots that never move into production. The answer wasn’t better models or more data—it was leadership, trust and the institutional foundations that allow AI to deliver better outcomes for residents.

What This Conversation Revealed

1. AI works best when governments organize around residents—not agencies.

The panel illustrated how fragmented systems create unnecessary friction for residents. In New York City, where childcare applications can involve three separate agencies, connecting existing data creates the opportunity to make services more proactive and significantly easier to access.

2. Scaling AI requires champions at every level.

One message came through consistently: executive sponsorship alone isn't enough. Neither is an enthusiastic department lead. Successful AI initiatives need buy-in at both the strategic and operational levels to move from pilot to implementation.

3. Trust starts long before deployment.

Whether discussing public confidence, responsible data use or vendor relationships, the panel returned to the same principle: AI should strengthen human decision-making—not replace it. Building that trust requires governance as much as technology.

Key Highlights

New York City showed the scale. More than eight million residents and roughly one million buildings—but no single view of either. A basic childcare application still spans three agencies.
Connected data changes the resident experience. A current pilot is securely connecting existing records so residents apply once while eligibility is inferred using existing housing or financial-assistance data.
Sponsorship has to exist at multiple levels. A departmental champion without executive backing rarely succeeds. Executive enthusiasm without operational ownership stalls just as quickly.
Governance is learned through real pilots. One county described renegotiating a vendor agreement after realizing existing terms didn't adequately protect how county data could be used.
Procurement remains the bottleneck. While pilots move quickly, outdated procurement frameworks often prevent successful projects from scaling. Cross-county collaboration and pilot-based procurement emerged as practical ways to bridge that gap.
Start with people, purpose and impact. The moderator's closing message was simple: begin with the resident and the problem you're trying to solve—not the technology.

By the Numbers

0 M+

residents served by NYC agencies

0 M

buildings in New York City

agencies for one childcare application

levels of sponsorship needed

In their words

“We don’t think AI should replace judgment. It should sharpen it, so that human decisions are sharper and more informed.”

April Rai, President & CEO, COMTO

Why it Matters

For years, conversations about AI have focused on what the technology can do. This discussion pointed to a different question: what kind of institution is needed to use AI well?

Across New York, Baltimore and Miami-Dade, the barriers weren’t model performance or access to data. They were questions of ownership, governance and trust. Who has the authority to act on an AI recommendation? How do agencies share data responsibly? How do procurement models evolve quickly enough to move successful pilots into long-term deployment? These are leadership challenges as much as technical ones—and increasingly, they’re what determine whether AI creates lasting public value.

That lesson extends well beyond government. As AI becomes embedded across critical infrastructure and operational systems, competitive advantage will come less from deploying new models and more from building organizations that can govern, adopt and scale them responsibly. The institutions that lead won’t necessarily be those with the most sophisticated AI. They’ll be the ones with the strongest foundations: trusted data, modern governance, aligned leadership and technology partners committed to building capability—not just delivering software.

Ultimately, the session reinforced a principle that echoed throughout Evolve: real impact doesn’t begin with AI. It begins with a clearly defined problem, the right institutional foundations and a commitment to improving outcomes for the people those systems are designed to serve.