Skip to main content

City Operations Through AI

Cities are adopting automated processes and operations (orchestrated by a city platform) and following data-driven planning approaches

In ancient Rome in the first century AD, the invention of aqueducts was critical for population growth. In the late nineteenth century skyscrapers in Chicago were important for managing land scarcity. Many other technologies and solutions have contributed to the foundation and development of vibrant cities. Artificial intelligence (AI) is now emerging as an essential part of how cities work.

Machines run 24/7, and there are operations and tasks that cities perform that will become increasingly smart and powered by process automation and artificial intelligence. AI will contribute to the optimisation of operational efficiencies, benefiting city managers and ultimately citizens through reshaped service delivery. Gartner predicted that by 2021 30% of city government service interactions would be fulfilled and/or completed, at least in part, through an AI-powered conversational channel.1 But the investment in AI is broader. 66 per cent of 167 cities inquired for ESI Thoughtlab study are investing heavily in AI and 80 per cent will do so over the next three years. North American cities (83 per cent) and small cities (74 per cent) lead in the use of AI.2

While chat assistants are currently among the most common solutions powered by AI, cities will evolve to have digital platforms as ’city brains’, where all urban activity is orchestrated and operated, providing a holistic view of the city, allowing for events correlation, fast and assertive ‘root cause’ analysis, predictive analysis (through machine learning) and incident management, and providing operational insights through visualisation. If the behaviour of almost every citizen is registered through anonymised data, and 5G technology enables cities to become huge connected ecosystems, it will be of paramount importance to maximize data value and improve planning and decision-making using AI and data analytics, on the way to a cognitive city. Gartner predicted in 2019 that a city platform will be a mature smart city solution in five to ten years’ time, when it is expected that one to five per cent of cities will be using a city platform to manage their operations.3

With a clear vision, proper infrastructure and data governance in place, cities should be expected to embrace digital transformation and leverage cloud computing and the Internet of Things, design new operating models that foster integration between inter-dependent departmental services, and automate intelligent operations further using AI – fostering better quality of services, and greater efficiency and effectiveness.

But cities can go even further. We see cities like Dublin and Singapore, among others, creating a Digital Twin – a dynamic digital replica of their physical assets and environments and their interdependencies – for urban planning purposes, and using machine learning to predict future events and trends. A Digital Twin can be used for example to provide support for day-to-day operations, to simulate a natural disaster and its potential impact on the city, or to evaluate the flow of breezes that cool the city and the trees to ensure shade in streets and parks. With the evolution of new technologies with higher processing capabilities (namely, fast problems root-cause analysis identification). Digital Twins will become increasingly powerful in enabling data-driven decisions, and will have a high adoption rate among city governments, with a promise of turning cities more resilient.4 ABI research has predicted that by 2025 the number of Digital Twins will exceed 500.5 And ESI ThoughtLab predicts that the percentage of cities making large investments will increase the most for Digital Twins, rising from 11 per cent in 2021 to 31 per cent in three years’ time—an increase of almost 300%.6

“What we have been working on is the transformation of data into relevant information for strategic decisions that we can make. This will improve immensely the governance and the efficiency of the city and ultimately the transparency of the decisions made by politicians or by public authorities.”

-Rui Moreira, Mayor of Porto

Quicker responses and better services: Leveraging big data analytics and machine learning enables a city to understand better what is happening and to adapt to a continuously changing environment, thus allowing for faster responses to new challenges. AI-enabled operations take data from all sensors and devices, so that the city can prevent faults or breakdowns, or identify a fault the moment the system goes down and put it right faster and automatically.7 Additionally, the city assesses its data and is always learning and responding to the needs and changing habits of their stakeholders (for example with a new bus lane or bicycle lane) and also receives service requests and suggestions from them. In the aftermath of the COVID-19 outbreak, 40 per cent of cities have admitted that timely access to data and advanced analytics are crucial for running a city.8

Safer and secure cities: Data from connected devices and AI-driven applications are used with analytics and image processing to understand what is going on in a city. Predictive tools can be used for example to help identify potential locations and times for certain crimes,9 and to support responses by the emergency and law enforcement agencies. For example, traffic intelligence and identification counts and predicts vehicle counts and flows, detects vehicle moving in wrong direction, identifies vehicles of interest, and so on. Crowd monitoring enables the city to notify the police department when crowd numbers reach threshold limits or vary significantly from predictions. For example the crime rate in Surat, India fell by 27 per cent after the implementation of AI-based safety measures.10

Efficient cities: AI in smart cities enables automation of municipal activities and operations on a large scale, reducing the duplication of efforts and improving effectiveness. It transforms the way in which cities operate and deliver public services, creating efficiencies and finding synergies. For example, Seoul established an integrated public transport system that uses smart cameras in subways to obtain information on passenger volumes and adjust the speed and frequency of trains in real time accordingly. It also installed sensors that monitor train components to prevent failures before they occur.11 São Paulo has developed a solution for estimating and predicting air quality using AI and big data analytics, using data from the mobile network complemented with data from weather, traffic and pollution sensors. This helps calculate pollution levels 24 - 48 hours in advance, allowing local governments to take preventive actions.12

Higher touchpoints between government and citizens: AI enables tailored and personalised liaison between local government and residents. The most effective policies can be developed applying AI to feedback from citizens. For example in North Carolina government offices use AI chatbots to speed up the process of responding to residents’ questions.13

Better disaster management and long-term planning: Another benefit of the application of AI and machine learning to city operations is in supporting short-, medium- and long-term planning. By connecting data from different sources, namely agencies, citizens, businesses, tourists, etc., the city urban planner can better identify trends and predict future needs and changing habits. City leaders can take data-driven decisions such as where to build a new school, or reinforce the bus network, or whether to open a new health centre in a district where the population is aging. That information is of paramount importance, not just for city planning; it has also value for businesses, enabling a better balance between demand and supply.

Start with data strategy and governance: Data governance and transparency are particularly important for cities that adopt AI solutions. Stakeholders within cities need to be made aware of how their data is going to be used – and for what purposes – so that people can trust the system. It requires an adjustment to the current city governance to make sure it entails a change of approach to data-driven decision-making - and eventually to an automated and integrated operation centre. A city must ensure the transparent exchange of quality, real-time, open data, and the ability to enrich the data – through monetisation mechanisms, a clearing house or blockchain for instance. Even if the city decides that some of the data is free, control mechanisms must be but in place to control abuse. Without this governance, trust in a city’s data marketplace (and its accuracy) will be fragile. Data governance models should build trust into their systems for data collection, privacy and data exposure, as it is key for political and public support. According to Gartner, by 2023 30 per cent of smart city initiatives will lose public support and be discontinued for lack of integrated services and data analysis.14

Be aware of privacy issues and stimulate a culture of trust: While the use of data can contribute to better delivery of services, privacy is a concern that must be properly addressed. Cities must respect data protection and security legislation and ensure proper use of personal data, in order to win and retain public confidence.

Ensure data standards and interoperability: It is crucial to maintain data standards and interoperability within the city, to facilitate seamless integration and analysis. Standardised methodologies like ISO 37120 and commercial data orchestrators facilitate that interoperability. Data integration would benefit from the existence of an API portal in the city – to protect the city’s digital platforms (or ’city brains’) and sub-systems against threats, vulnerabilities, and with controls over access with single sign-on and identity management, providing end to-end security.

Avoid algorithmic bias: All AI systems use algorithms, which may be biased in the way they function. It is particularly important that algorithms should not be biased in a way that deepens inequalities (for instance, between racial or ethnic groups). Having a diverse team working with data can mitigate this.

Prepare the right skill set among the city workforce: Cities will have to provide many government workers with effective short-term training programmes and lifelong learning to help them adapt to AI. Existing educational programmes will also have to be revamped to provide skills in AI to individuals entering the workforce in the future. 15

Follow a citizen-focused approach to operations: Putting citizens, local businesses and visitors at the centre when designing city operations is the way to delivering better city services.

Cascais, Portugal

Cascais, Portugal, is a coastal resort town with a population of 211,000 that attracts more than 1.2 million tourists a year and aims to be “the best to live for a day or a lifetime”. To drive efficiencies in infrastructure, transport, public safety and other services, the city has a mission to “test innovative solutions capable of being scaled.” It has developed a large portfolio of technology-based services ranging from energy-efficient buildings to remote payments for parking.

However, Cascais faced challenges as it evolved its ecosystem and implemented new initiatives. One of the biggest obstacles was the lack of a unified vision across 12 municipal domains, ranging from health and education to energy and public infrastructure. To address this problem, in 2018 Cascais developed a managed services digital command centre, C2, to give it a holistic and integrated approach to the management of city operations in a multidisciplinary room. The solution was powered by Deloitte’s smart place operating system, CitySynergy.

Cascais redefined the city’s operating model by integrating data and processes from each municipal vertical domain instead of dealing with each in separate silos. Integration increased the quality of services to citizens and achieved savings based on higher effectiveness and efficiency.

The city platform now provides 15 smart initiatives (including citizen connection websites and a citizen engagement app) with integrated maps with assets and dependencies, online dashboards, customised reports and a Digital Twin. It supports management of an ecosystem of more than 30 service partners, enables predictive management through event correlation and data analytics, and facilitates decision-making and urban planning. “A command centre with predictive capabilities, to try to anticipate the future, that's what citizens want in the future”, says Miguel Pinto-Luz, the Deputy Mayor of Cascais.

C2 has helped improve operations, increase efficiencies and cut costs. For example, Cascais has implemented a smart waste management system that is expected to reduce journeys along routes by 180,000 kilometres and carbon dioxide emissions by 350 tons per year, producing savings of around EUR 600,000 annually.16 By integrating real-time traffic and road condition data, the system not only optimises routes but also identifies the best times for garbage collection, potentially reducing operating costs by up to 40 per cent and boosting energy savings by up to 20 per cent. Cascais has also improved citizen satisfaction levels and achieved 20 to 30 per cent for energy savings, and 30 per cent reduction in water consumption. The city is proud of having signed Service Level Agreements with its citizens and with the outcomes this has brought.

With its efficiencies, Cascais can allocate resources more effectively and attract new businesses, residents and universities, making it the most dynamic and forward-looking city in Portugal. More importantly, the model developed by Cascais could be replicated by other cities around the world. 17, 18

Vienna, Austria

Vienna was one of the first cities in the world to publish open government data in 2011, but its platform VeroCity took open data to a new level. Its data aggregation and analysis capabilities are based on the European Commission’s Context Broker building block, which can sort through data of all sorts and sources.19

The Context Broker allows the platform to offer real-time visual information that caters to all stakeholders in the city. The platform can facilitate day-to-day activities, such as urban mobility, environmental monitoring, urban infrastructure management, energy efficiency improvement and much more. The platform provides access to visualised information for users, avoiding the need to work through details in the raw data. This enables the city to deliver transparency in monitoring and benchmarking, while promoting participation by its citizens.

The city has also launched WienBot, a chatbot that provides answers to a range of user questions while also continuously learning from its ‘conversations’. This ability to capture most frequently-asked questions or used keywords, enables the chatbot to suggest questions in advance. Currently WienBot answers questions on the 250 most frequently-accessed contents of the City of Vienna’s official website www.wien.at. It also suggests other useful city services that might help users. The list of questions was updated recently in response to the COVID-19 pandemic.

As a result of these efforts towards technological management of city operations, for the second year in a row, in 2019 Vienna ranked first in the Smart City Strategy Index.20, 21

Hong Kong

Hong Kong is constantly augmenting the use of AI in the government and public sector. A priority is continuous improvement in the management of the city’s services. For example, the city is planning to deploy chatbots to use historical data to respond to citizens’ complaints and answer questions.

The city also plans to use AI in traffic management. The city already collects real-time traffic data on speeds and volumes via sensors across 80 per cent of major routes, to reduce congestion around the city.

The city also has sensors collecting data on landslides, pollution and water levels, so that it is better prepared for disasters, and it also uses sensors to monitor energy use.

A third of Hong Kong’s population will be aged 65 or above in 20 years’ time. The city plans to use robotics to support the elderly and assist care providers. Further, the hospitals in Hong Kong have deployed AI to schedule weekly tasks for thousands of nurses.

Visa applications are inspected by AI to prevent errors and misconduct. The city also plans to transform the ‘digital persona’ and use AI to create e-Identity for every individual. This will ensure that individuals with a trusted authentication can gain trouble-free access to private and public services online.23, 24

“I tend to not really like the label ‘smart cities’, because I don't think that there are dumb cities out there. There are cities that just need to harness technology to serve their citizens’ needs better, and that will vary from one city to another, depending on their needs, depending on their ability to leapfrog existing older technologies, depending on their capacity.”

-Sameh Wahba, Global Director of Urban, Disaster Risk Management, Resilience and Land Global Practice at the World Bank

Video Interviews

Podcasts

End Notes 

 
  1. Gartner, quoted in B-CITI: Artificial Intelligence at the service of smart cities v2.0. (2019)
  2. ESI ThoughtLab: Smart City solutions in a riskier world. (2021)
  3. Gartner: Hype Cycle for Smart City Technologies and Solutions. (2019)
  4. Open & Agile Smart Cities: Cities & Digital Twins: From Hype to Reality. (2020)
  5. ABI Research, quoted in Cities Today: COVID-19 expected to drive adoption of city digital twins. (2021)
  6. ESI ThoughtLab: Smart City solutions in a riskier world. (2021)
  7. Enterprise IoT Insights: Las Vegas applies AI to smart city operations to detect faults and outages (2018)
  8. ESI ThoughtLab: Smart City solutions in a riskier world. (2021)
  9. CIO: AI Conversations: The Rise of the Digital City (2020)
  10. SmartCity.Press: The Impact Of Artificial Intelligence Over Smart Cities. (2017)
  11. Terminus Group: AI CITY Will Pave the Way For the Cities of Tomorrow. (2020)
  12. ITU: Smart sustainable cities. (2019)
  13. ICMA: Using Artificial Intelligence as a Tool for Your Local Government (2019)
  14. Gartner: Establish an Urban Data Exchange for Smart Cities. (2020)
  15. MIT Sloan Management Review: How Cities Should Prepare for Artificial Intelligence. (2019)
  16. Canal Cascais: Smart Waste Management in Cascais benchmark in Portugal and abroad. (2018)
  17. Deloitte: Cascais; Command center approach to drive efficiencies at Cascais. (2019)
  18. The Wall Street Journal: How a Portuguese City Drives Efficiencies With Innovation. (2018)
  19. Vienna’s Open Data is published on the website ‘Open Data Osterreich’ and also at ´Open Government Data´.
  20. The European Commission's 100 Intelligent Cities Challenge: WienBot (interactive messaging service).
  21. European Commission: The number one smart city in the world uses CEF Context Broker to effectively manage Big Data. (2019)
  22. OSIsoft: Using data to predict and mitigate floods. (2015)
  23. GovInsider: Exclusive: Hong Kong’s vision for Artificial Intelligence.

You may access the links to these sources, where available, on page 148 of the Urban Future with a Purpose study.

Did you find this useful?

Thanks for your feedback

If you would like to help improve Deloitte.com further, please complete a 3-minute survey