People first, then technology: Lessons from cities leading on AI

Around the world, governments are putting intelligence to work across the services residents rely on every day. They're building cities that stay resilient when conditions change, grounding AI in information people can trust, and keeping their data and critical systems under their own control.

Jakarta, one of the world's most flood-prone cities, uses AI to forecast flood risk up to six hours ahead, giving authorities time to close floodgates, activate pumps, and alert residents before the water rises.

Aberdeen City Council, home to Scotland's first local-authority virtual assistant, worked with Microsoft to upgrade the tool with generative AI grounded in council knowledge. Residents now get specific answers around the clock, and soon in more of the languages their community speaks.

In Raleigh, North Carolina, the city worked with Microsoft, NVIDIA, and Esri to build an AI-powered digital twin of its traffic in six weeks. It estimates that trimming 4.5 seconds of delay at 10 signals on one of its busiest corridors could save commuters $9.7 million in time and travel costs. Three different challenges, one pattern: governments are putting AI to work where residents feel it most, and new research shows how far that momentum has spread.

Measure the momentum: What 200+ cities reveal about AI adoption

ThoughtLab's new research, Building a People-First, AI-Powered City, benchmarked more than 200 cities worldwide, surveying 525 senior officials and interviewing more than 50 urban leaders. The findings show how widely cities already use AI:

• 80% of cities already use AI for data management and 80% for infrastructure management.
• 77% apply AI to cybersecurity and data privacy.
• 71% apply AI to public safety and emergency readiness.
• 69% actively or widely use AI in digital citizen services.

Lead with people: How top cities more than double their AI returns

The cities pulling ahead share one trait: they start with people, not technology. ThoughtLab designates Leaders as the top tier of its AI maturity model. Their biggest edge is human-centered design, where they score more than twice as high as other cities.

Leaders begin with residents' needs, rights, and concerns, then choose the AI that delivers public value. Every Leader builds cybersecurity into the AI lifecycle, and nearly all keep sovereign control of sensitive data.

Compared with other cities, Leaders report:

• A 99% average return on AI investments, versus 42%.
• Payback 11 months sooner: 38 months versus 49.
• A 1.8 times greater likelihood of building public trust.

Next, three Leader cities show what this looks like in practice. Any city can adapt these approaches for its own people, policies, and systems.

TAMM, Arabic for "consider it done," is Abu Dhabi's government services platform, powered by Microsoft. It brings more than 1,200 services to 4.7 million users. Customer satisfaction has climbed to 94.5%, up from 78% before AI was integrated, and service costs have dropped 30%.

TAMM AutoGov uses agentic AI to support proactive trade license and disability card renewals before they expire, reducing the need for residents to initiate the process. "A government service that works well is one that reaches people before they have to ask for it-before the deadline, before the disruption," says His Excellency Dr. Mohamed Al Askar, Director General of TAMM. The United Nations-backed WSIS Prizes named TAMM the world's best e-Government project in 2025, and in October it won the Government category for EMEA in the Gartner Eye on Innovation Awards 2026.

Build for reuse: Tampere

Two Tampere developers and a Microsoft team built Tampere VIP, an agentic city experience assistant piloted at Finland's national Eurovision selection. People ask by voice or text for places to eat, events, and activities, and several AI agents share the work, each drawing only on predefined, reliable data. Swap the agents and their data, and the same platform serves a new purpose without a rebuild. In June, Gothenburg, Los Angeles, Turku, and Valencia joined Tampere to explore the approach.

Free dispatchers for emergencies: Munich

Munich Fire Department's control center handles up to a million calls a year, including about 120,000 non-urgent patient transport requests. Many of the nurses who book those rides speak German as a second or third language. Working with Microsoft, firefighters built an AI voice operator, now in testing, that takes those bookings and keeps dispatchers focused on emergencies. If it detects uncertainty or a possible emergency, it hands the call to a person, and dispatchers make every decision. "The bot helps but the human saves," says Mathias Duensing, head of IT architecture at the Munich Fire Department.

Design for trust from the start

How do cities turn promising pilots into trusted services? ThoughtLab found that the most successful cities disclose where AI is used, keep humans accountable for high-impact decisions, protect privacy, and show residents concrete benefits.

City leaders need to know where their data resides, who can operate their systems, and whether critical services keep running when conditions change. Microsoft Sovereign Cloud and the latest digital sovereignty and Azure Enclave announcements help address those questions:

• Foundry Local, becoming generally available at Microsoft Ignite, runs AI models on Azure Local in connected and disconnected environments that customers operate and control.
• Azure Enclave, now generally available, gives governments isolated, policy-governed environments for working on sensitive data with partners, research institutions, and other agencies.

See it in action at Smart City Expo World Congress 2026

From November 3 to 5, Microsoft will be at Fira Barcelona Gran Via, Hall 3, Stand 3D51, with a booth built around three themes: resilient cities, trusted intelligence, and sovereignty you control.

Explore frontier demos across three city priorities

• Personalized government services (Capgemini, Accenture, Zammo.ai, and Microsoft Frontier Company's city experience assistant, featuring Tampere).
• Workforce decision-making (3AM Innovations, Cyclomedia, Bentley Systems, and EY).
• Resilient government operations (KPMG and Tomorrow.io, Scenexus, Esri, Microsoft Security, and Microsoft Sovereign Cloud).

Build and test an agent in the Agent Experience Zone

Pick one of six city use cases, from permitting and housing eligibility to shift handoffs and disaster response. Stress-test it in the Trust Lab: Can we trust the source? Is permission clear? When must a person take over? Then build your own in a guided quick build.

Light up your city

Add your city to a world map that lights up as attendees build throughout the week, and take home a personalized city sticker designed with AI.

Hear from cities in the Microsoft Theater

There will be 20+ sessions that include:

• AI & Cities with ThoughtLab + City of Manchester (November 3, 3:00 PM CET).
• Inside Abu Dhabi's Exportable AI-Native Government Blueprint (November 4, 11:00 AM CET).
• From First Agent to Citywide Impact: Scaling Tampere's Smart City Model (November 4, 2:00 PM CET).
• Trust by Design: Women Leading AI Deployment in Government, with the mayor of Mount Vernon, New York (November 4, 3:00 PM CET).

Catch Microsoft leaders on stage

On November 3, hear from:

• Kees Hertogh, Vice President of Global Industry Marketing, with Tampere Deputy Mayor Jaakko Mustakallio (11:25 AM CET, Main Stage): Trusted Intelligence in Action: Tampere's City Experience Assistant with Microsoft. Learn what makes an AI-supported city service useful, trusted, and accountable.
• Kate Maxwell, General Manager, Worldwide Government, Defense & Intelligence Industry (12:05 PM CET, AI X Cities Stage): The City as a Living System: Navigating the Era of Agentic AI. Learn where agentic AI already works in real-world settings and what it means for city services.
• John Doyle, Head of Public Sector Industry Marketing (10:15 CET, AI X Cities Stage): Augment the Workforce: How Agents Multiply the City's People. See how agents capture institutional knowledge and absorb routine casework while people decide and handle the exceptions.

Turn people-first AI momentum into action

Cities are already proving what people-first AI can deliver: services that reach residents sooner, staff freed for the work that matters most, and data that stays under their control.

Join us in Barcelona to see what comes next, explore ThoughtLab's research to plan your city's next step, and learn more about Microsoft for government.