Connecting Data and Services to Build People-Centred Digital Cities

In Brief

The core problem: Most digital city initiatives fail not because of technology, but because of fragmented data, poor community relationships, and insufficient digital literacy among residents.

What works: Cities that succeed treat citizens as co-designers, not end-users. They connect data across departments, build trust before deploying systems, and measure outcomes that matter to people.

Implementation reality: A people-centred digital city requires three things: observable data pipelines, clear ownership of failures, and rollback plans for every service launch.

The governance gap: UN-Habitat's international guidelines on people-centred smart cities, expected in 2025, will provide the first global framework for ethical digital urban infrastructure.

This conversation continues at Human x AI Europe on May 19 in Vienna, where policymakers, technologists, and city leaders will work through the practical challenges of building digital infrastructure that actually serves people. Join the room where Europe's future gets built.

The Model Is the Easy Part

Every city wants to be smart. Few want to do the work that makes smart actually useful.

The pattern is familiar: a municipality procures a platform, launches a pilot, holds a press conference, and then watches adoption flatline. Six months later, the dashboard shows impressive data visualizations that nobody uses, the chatbot answers questions nobody asked, and the app sits at 2.3 stars in the app store with reviews like doesn't work and can't find my street.

This isn't a technology problem. According to the Aspen Institute's research on digital infrastructure, the delivery of effective digital services at the local level has been hamstrung by three primary obstacles: an absence of data, poor community-government relationships, and insufficient digital literacy among key populations. At their core, these failures all have one thing in common: they designed for the community but not with the community.

That last sentence should be printed on every smart city RFP.

What People-Centred Actually Means in Practice

The phrase people-centred gets thrown around in procurement documents like confetti. Strip away the marketing language and ask: what does it mean operationally?

UN-Habitat's People-Centred Smart Cities programme offers a working definition: deployment of technology and innovation used to ensure sustainability, inclusivity, prosperity, and human rights in cities. The programme supports national and local governments with their digital transition, applying a multi-level governance strategy to help build skills and capabilities to develop, procure, and effectively use digital technologies in an ethical and inclusive way.

That's the principle. Here's the implementation checklist:

Before any digital service launch, answer these questions:

• Who gets paged when this breaks at 2 AM on a Saturday?
• What does good enough look like for the first 90 days?
• How do residents who lack smartphones or reliable internet access the same service?
• What's the rollback plan if adoption is below 10% after six months?
• Which community groups were involved in defining the problem, not just testing the solution?

If the project team can't answer all five, the project isn't ready to ship.

The Data Problem Nobody Wants to Talk About

Research from the University of Texas documented how data and information inadequacy hampered India's pandemic response across multiple dimensions: delayed threat assessment, inefficient track-and-trace, inability to locate vulnerable populations, and failure to deliver financial aid to unbanked citizens. The same patterns appear in cities worldwide, just with different failure modes.

The uncomfortable truth: most cities don't know what data they have, where it lives, or who owns it.

A city might have excellent property records in one system, utility data in another, and social services information in a third. None of them talk to each other. When a crisis hits, staff resort to spreadsheets and phone calls because the integrated platform requires three weeks of IT tickets to generate a cross-departmental report.

The National League of Cities documented how cities are addressing this: Plant City, Florida, used integrated ERP (Enterprise Resource Planning) and financial systems to expedite maximum FEMA reimbursements after Hurricane Irma. Mobile, Alabama, saved more than $500,000 annually simply by linking police and court records, shrinking police department paperwork by 98 percent.

These aren't moonshot projects. They're plumbing. And plumbing is what makes cities work.

The Governance Framework That's Coming

In June 2023, 193 countries requested UN-Habitat to develop international guidelines on people-centred smart cities through Resolution HSP/HA.2/Res.1. According to UN-Habitat's programme documentation, these guidelines will serve as a non-binding framework for developing national and local smart city regulations, plans, and strategies, to ensure that digital urban infrastructure and data contribute to making cities and human settlements sustainable, inclusive, prosperous, and respectful of human rights.

The guidelines development process is expected to be finalized in 2025. For cities planning digital infrastructure investments now, this creates a decision point: build to current ad-hoc standards and retrofit later, or anticipate the framework and build compliance in from the start.

The second option costs more upfront. The first option costs more in total.

What Winning Cities Actually Do

The 2025 Digital Cities Survey from the Center for Digital Government identified patterns among first-place winners across population categories. The common threads: community input driving innovation, data governance supporting AI applications, and upgraded constituent interfaces.

South Bend, Indiana, ran a generative AI roadshow where staff visited different city offices to brainstorm AI use cases. More importantly, they used their annual data inventory and governance process to ensure data was AI-ready before deploying any models. As their Chief Innovation Officer noted, AI is only as good as the underlying data.

Scottsdale, Arizona, deployed Microsoft Copilot to over 1,000 users and launched pilot projects in finance, HR, permitting, and public safety operations. Their CIO described the approach: We start with a problem. What are the challenges or problems, the services that we want to enhance? And from there we co-create solutions.

Tamarac, Florida, implemented an Amazon Connect omnichannel call center where just over 40 percent of call traffic was fielded by AI, with half those callers getting answers without connecting to a human. The system works across phone, SMS, web chat, and video, providing consistent service regardless of channel.

The Implementation Template

For teams building people-centred digital city services, here's a starting framework:

Phase 1: Discovery (4-8 weeks)

• Map existing data sources across departments
• Identify which community groups are underserved by current digital services
• Document current failure modes and workarounds staff use daily
• Establish baseline metrics for service delivery times and satisfaction

Phase 2: Co-Design (6-12 weeks)

• Recruit community members as co-designers, not just testers
• Build paper prototypes before writing code
• Define good enough thresholds with stakeholders
• Establish ownership: who gets paged when it breaks?

Phase 3: Pilot (8-16 weeks)

• Deploy to a limited population with clear success criteria
• Monitor for drift weekly, not quarterly
• Collect qualitative feedback alongside quantitative metrics
• Document what breaks and why

Phase 4: Scale or Kill (4 weeks)

• If metrics meet thresholds, expand gradually
• If metrics don't meet thresholds, either pivot or shut down
• Write the postmortem either way

The timeline matters less than the sequence. Skip co-design and the pilot will fail. Skip the postmortem and the next project will repeat the same mistakes.

The Question That Matters

Before the next smart city initiative gets funded, ask one question: What happens when this fails?

Not if. When.

Every system fails eventually. The difference between a people-centred digital city and a technology showcase is whether the failure mode was designed for, whether someone owns the recovery, and whether residents can still access services when the platform is down.

Build the rollback plan before the launch plan. That's what people-centred actually means.