AI-leading cities report more than double peer returns
A global survey of 525 officials across 215 cities finds AI leaders average a 96% return on investment and recoup spending 10 months faster, even as 69% cite trust gaps.

Why it matters
- ThoughtLab surveyed 525 senior officials across 215 cities and reviewed 363 city AI plans.
- AI-leading cities reported 96% average ROI versus 42% for peers and recouped spending 10 months faster.
- Only 27% of respondents said their city had an integrated, citywide AI strategy.
- Honolulu cut permitting review cycles by 58% and applicant waiting times by 40.5 days on average.
- Among the 37 leader cities, 95% pursued sovereign-data policies and 89% expected AI to become a key economic driver.
Cities classified as artificial intelligence (AI) leaders by ThoughtLab report an average return on investment (ROI) of 96%, more than double the 42% reported by other municipalities. They recoup spending 10 months faster, according to the global study published on November 17, 2026.
The gap carries policy stakes because ThoughtLab links stronger returns to cities deploying AI at scale rather than limiting programs to tests. Its research associates those returns with operating models that connect AI use with social, economic and environmental goals.
ThoughtLab based the study on a survey of 525 senior officials across 215 cities. It also reviewed 363 city AI plans and held discussions and interviews with government, business and academic leaders.
The study points to a small group of cities that have moved beyond pilot schemes. Its examples focus on everyday municipal functions rather than experimental consumer tools, according to ThoughtLab.
Where does AI deliver measurable municipal gains?
ThoughtLab documents gains across permitting, transport, utilities, emergency response and traffic management.
- Honolulu: An AI-based plan review tool cut permitting review cycles by 58%. ThoughtLab reports an average reduction of 40.5 days in applicants’ waiting times.
- New York City: The Metropolitan Transportation Authority (MTA) uses predictive AI to identify bus maintenance issues before they disrupt services. The system has cut material costs by 24%.
- Gold Coast, Australia: AI-based water monitoring and predictive maintenance detect leaks. ThoughtLab reports returns of 60% to 80% from those uses.
- Seoul: An AI emergency call chatbot gathers information and helps operators prioritize urgent reports.
- Kigali: Traffic management in Kigali has reduced travel times by as much as 30%.
Across these cases, city administrations apply AI to public safety, transport, infrastructure, permitting and utility management. More advanced cities use the technology not only to automate administrative work but also to coordinate decisions across departments, according to ThoughtLab.
Which cities have a joined-up AI strategy?
Only 27% of survey respondents said their city had an integrated, citywide AI strategy. The governance gap is stark.
ThoughtLab associates the strongest returns with cities that combine governance, staff training, data controls and service design around resident needs.
Among the cities ThoughtLab classifies as leaders, 92% train employees in the ethical and responsible use of AI. Seventy-nine percent build fairness, transparency and inclusion into planning. The same 79% have formal transparency guidelines for AI-enabled services.
Those practices sit inside the operating model that ThoughtLab links to stronger returns. The report treats staff preparation, public data controls and service design as connected parts of AI governance rather than separate initiatives.
ThoughtLab’s model places cities into four groups: Starters, Implementers, Accelerators and Leaders. The leader category contains 37 cities across regions and sizes.
Why does public trust constrain deployment?
Public confidence remains a major obstacle. Sixty-nine percent of surveyed cities said trust and public acceptance were among their biggest challenges, according to ThoughtLab. The report treats that gap as a political and administrative risk for wider deployment.
The report also frames human oversight as an unsettled governance choice as city administrations decide where human review should remain. ThoughtLab links oversight to the same operating model that supports scaled services and higher returns.
Leading cities are more likely to treat control of public data as a policy issue. Among the 37 leaders, 95% said they are actively implementing policies to maintain sovereign control over sensitive city and citizen data. Meanwhile, 89% said AI will become a key competitive and economic driver for their city.
Geoffrey Urbach, Product and Partnerships Lead, Emerging Technology, San Antonio, said:
“Leaders who treat these concerns as a communication challenge rather than a governance issue will erode the trust they're trying to build.”
Urbach’s warning aligns with the survey’s governance data: training, transparency and data control recur among leaders, while trust remains a citywide constraint.
Who are the 37 AI-leading cities?
ThoughtLab names 13 cities from the 37-city leader category. The list spans four regional groupings:
- Europe: Amsterdam, Barcelona and Dublin.
- Asia-Pacific: Melbourne, Seoul and Singapore.
- Americas: Austin, Buenos Aires, Montreal and Seattle.
- Middle East and Africa: Abu Dhabi, Dubai and Kigali.
ThoughtLab presents the full 37-city group as spanning both regions and sizes.
What does a people-first approach change?
ThoughtLab’s central argument is that city governments gain more from AI when they design services around residents rather than around the technology itself. The study links that approach not only to higher financial returns but also to stronger outcomes in trust, social equity and sustainability.
“Our analysis reveals a fundamental shift in how cities think about AI,” said Lou Celi, Chief Executive Officer, ThoughtLab.
“City leaders see AI as a way to reinvent how government processes work, how services are delivered, and how cities anticipate and respond to residents' needs. What's most encouraging is that they are using AI in a way that strengthens public value, trust, and social equity.”
Celi’s account restates ThoughtLab’s people-first thesis. AI changes how government processes work, how services reach residents and how administrations anticipate demand.
Nicole Coughlin, Chief Information Officer, Cary, North Carolina, described the operational effect:
“AI gives our employees something incredibly valuable: time,” said Coughlin.
“By reducing repetitive administrative work and surfacing better insights, it allows public servants to focus on the work that matters most-solving complex problems, collaborating across departments, and serving people.”
Coughlin’s account matches ThoughtLab’s finding that more advanced cities automate repetitive administration while also coordinating decisions across departments. The measured benefit extends beyond time saved, according to the study, to public value, trust and social equity.
As more administrations move from pilots into citywide services, ThoughtLab’s evidence makes institutional capacity the next constraint on performance. Only 27% of respondents report an integrated citywide strategy, while 69% still identify trust and public acceptance as major obstacles.
For cities outside the 37-city leader group, the next test is whether governance, employee training, data controls and service design can develop as quickly as their deployments. The report’s 96% return figure gives leaders a benchmark, but not a guarantee for cities that scale without those foundations.
Source: itbrief.com.au
More from Marcus Bennett
Show full bio
Senior reporter covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
718 articles
Related articles
- Mamdani Says Feds Have Abdicated on AI, Urges New Yorkers to Push Back
- 51% of AI projects miss targets, yet 85% of firms raise budgets
- Camouflaged Data Centers: San Antonio Tests Whether Hiding AI's Footprint Helps
- FAA Expands AI Flight Planner SMART to All 22 En Route Centers
- Google DeepMind's PlanAssist Cuts Planning Officer Workload by 58%