Mostafa Othman, smart cities programme director, Honeywell, examines how AI is heralding a new urban intelligence where cities learn, adapt and optimise themselves in real time.
Technologies like AI, IoT and digital twins are endowing cities with a level of intelligence that would have been unheard of as recent as five years ago. Indeed, what passed as a smart city a few years ago is already starting to look passive by comparison. With more and more technology and intelligence being embedded in cities, the discussion is moving towards the concept of a cognitive city.
So how does this differ from a smart one? Historically, smart cities collect data via technologies such as the IoT and use a rules-based engine to interpret this data to carry out an action. The focus has been on automation and efficiency. For example, if a city wants to optimise street lighting to save energy but maintain safety, it can use such technologies to control when and how lights operate.
A cognitive city makes use of technologies such as artificial intelligence, machine learning and predictive analytics to create one that learns, adapts and optimises itself in real time. It delivers the same automation and efficiency gains but is more context-aware, self-improving and responsive to the needs of the citizens as well as those who run them. In short, they become active participants in the urban landscape.
Data underpins a cognitive city much like it does a smart one, but the difference is that AI enables it to learn from the data rather than just act on it. Crucially, it can correlate previous unrelated data points. These enhanced and expanded data processing capabilities also increase the relevance and power of the data across the different departments and layers of city operations. It informs how an action taken by one department or operator will impact another and helps to break down both the data and people silos that often stymie the workings of a smart city. A leaky pipe, for instance, isn’t just a water utility issue but also a problem for those managing traffic, safety and public works.
More personalised and predictive cities
AI is also helping cities to be more personalised and predictive when it comes to managing assets and delivering services to the community. Instead of waiting for a sensor to trigger an alert about a streetlight, AI can analyse historical and real-time data to forecast when one of its components might fail, enabling the city to move from reactive to proactive maintenance. Meanwhile, if an incident has taken place impacting transit, it could be used to inform and re-route both drivers and pedestrians.
The granular detail available means it could even tailor information to those with accessibility requirements. Moreover, AI doesn’t just support decision-making, it simulates it, enabling those who run cities to model outcomes of a proposed policy, such as the impact of a new bus route on congestion and emissions. Feedback from those who live and work in the city can also be factored in to further inform the decision-making process.
The IoT has enabled us to reach a high level of smartness and turned silent assets into those which communicate with us. But the context that connects the systems that run them across different platforms or departments has been missing. AI provides this vital piece in the jigsaw and shows that context can be truly life-saving.
For example, imagine a sensor detects a leak in one of the city’s water pipelines. That information currently goes only to the water department, which sends a field team to repair it. But the system doesn’t understand how this issue affects other departments or the wider city context. The road above the leak might be busy with cars – one of which could be an ambulance rushing a patient to the hospital. This is where AI can make a difference: by understanding the context across departments and even making or implementing decisions on its own.
The role of digital twins
Digital twins are emerging as another key strand to cognitive cities. They provide extensive monitoring capabilities – once again enabling a better understanding of context – but also bring a new level of visualisation. We can see the entire city and its buildings in 3D, almost as if we are walking through the streets or moving inside the buildings. This gives us an incredibly immersive way to understand everything happening across the city.
However, it’s not only about monitoring or visualisation. The most important benefit of the digital twin lies in its ability to simulate scenarios and empower decision-making. The simulation allows us to understand the potential impact of any decision in real time, through an actual visual representation rather than just a report. Those running cities can literally see how a change will affect people, infrastructure, and assets.
Of course, while the digital twin offers tremendous value, it also comes with a number of challenges that need to be addressed, as we’ve seen in various projects.
The first of these is around data. Building a complete digital twin of a city requires a huge amount of data, and many cities were not initially prepared for that. Often, the necessary data simply doesn’t exist or isn’t available in the right format or quality. Ensuring data completeness, accuracy, and consistency is crucial. In some cases, one department may hold one version of the data, while another has a completely different version. Maintaining consistent and reliable data at scale is therefore one of the most critical aspects.
The second challenge relates to infrastructure and storage. We are not just dealing with standard database entries, but with massive files such as BIM and CAD models. This means significantly higher storage requirements and, when running AI use cases, a need for GPUs at the server or data centre level. These factors increase both cost and the technical capabilities required, whether the systems are hosted on-premises or in the cloud. This also affects the maintenance costs of the overall system and infrastructure.
Finally, one of the most important challenges is people as using a digital twin, running simulations and leveraging insights usually requires new skills. That means capacity building and training are essential. Addressing the skills gap is key, because it doesn’t make sense to invest heavily in technology and infrastructure without also investing in the people who will operate and benefit from it.
| City challenges and filling AI gaps
As cities continue to evolve, they must juggle myriad challenges, from aging infrastructure and climate risks to workforce shortages and increasing demands for better services. Discover how AI can be embedded into urban environments to overcome these challenges by reading our Insight Report with Honeywell – Smart, safe, sustainable: overcoming city challenges using AI – and tune into the accompanying on-demand webinar to see use cases brought to life and key city questions answered. |
Cities Maturity Framework
At the moment, cognitive cities remain more aspiration than the norm and going forward, there are two distinct tracks: one related to cities themselves, and the other to technology.
From the city perspective, while technology continues to evolve rapidly, the actual rate of adoption and implementation in cities hasn’t yet caught up with this pace of innovation.
At Honeywell, we use the “Cities Maturity Framework”, which helps us assess a city’s level of technological maturity. We’ve identified four levels:
- Level 1 – Ad Hoc: systems operate in silos with no integration
- Level 2 – Developing: there is minimal integration between systems
- Level 3 – Integrated: systems are data-driven and interconnected
- Level 4 – AI-Driven: operations are predictive and autonomous.
Most cities today probably still fall somewhere between Level 2 and Level 3, while the technology itself has already reached Level 4. From a city standpoint, we are gradually moving toward more integrated frameworks and coordinated operations, and only then will we start to see broader adoption of AI-driven capabilities. This transition will take time – probably two years at least – because such decisions are complex and cannot be made overnight.
From the technology perspective, I believe we are moving strongly toward autonomous operations. This includes areas such as autonomous vehicles, autonomous waste management, and even self-operating infrastructure that can respond automatically to alarms, readings, and incoming data. While this autonomy may not cover every domain, it will become increasingly common in sectors like transportation, traffic management, and waste services.
I believe the future will be increasingly human-centred, with a focus on personalisation. For example, transportation systems might offer personalised ticketing based on individual needs, and healthcare services maybecome more tailored to each person’s specific data and context.
Cognitive cities represent the next evolutionary step in urban intelligence – moving beyond systems that simply automate tasks to ones that truly think, learn, and evolve. By fusing the intelligence of AI, the connectivity of IoT, and the insight of digital twins, cities can move beyond efficiency to achieve true awareness and foresight. But realising this vision will require continued investment in data quality, digital infrastructure, and above all, people – ensuring that those who design and operate these systems have the skills to harness their full potential.
When these elements come together, cities will no longer just be “smart”; they will be cognitive – living, adaptive environments that enhance the quality of life for all who inhabit them.
---
Autor(en)/Author(s): Mostafa Othman
Dieser Artikel ist neu veröffentlicht von / This article is republished from: Smart Cities World, 31.10.2025
Bitte besuchen Sie/Please visit:
