Article

Digital, human, inclusive – how to design a workplace for neurodivergent people

As we learn more about neurodiverse experiences, the question of what makes a stimulating and appealing workplace has never been more complex. Our new, machine learning powered methodology is helping to make design more inclusive and effective.

A group of people sitting at tables in Arup Sydney office

James Ward

Associate Director

Last updated: July 2025

Today, researchers believe that between 15-20% of us are in some way neurodivergent and therefore experience the world, learn or think differently to the majority. Some people value face-to-face communication, others depend on quiet, lower lit zones for concentrated effort. And within this wide-ranging group you’ll often find the most original thinkers or people with great contributions to make – if only they could find a workplace that’s right for them.

Today, successful businesses and organisations recognise that they need to employ a diverse range of talents, from different ages, genders, skillsets and yes, neurodiverse profiles. In essence, if a workplace doesn’t work for everyone, it doesn’t work properly at all.

In recent years, the question of what makes a stimulating, effective and appealing workplace has never been more complex. Developers and tenant organisations face a mix of various trends, preferences, technologies and financial priorities. Hybrid working. Collaborative working. Hot desking. And increasingly, a workplace that works for neurodivergent workers.

Navigating the modern workplace

To understand the issue, let’s picture a typical modern office. A brightly lit, noisy space. There might be over 300 coworkers on a single floor. Often these are open plan spaces – to create a shared, ‘buzzy’ working environment. By 9.45am there’s a hubbub of noise and movement, people on Zoom calls, or chatting across desks, or gathered around screens, discussing the latest thing. Maybe the company is in media or tech – so it’s a former warehouse – high ceilings provide bright light, but make for an echoey and sonically overwhelming experience at times. At one end of the room a few wooden ‘call pods’ have been installed – for those needing seclusion and quiet. But they are rather small and claustrophobic. Over time, this office has developed to reflect the business’s own messy growth path – it’s a jumble of decisions, possibilities and outcomes. For neurodivergent workers, there are so many challenges to working effectively in this kind of space.

Let’s be clear – for that 15-20% of the population who are neurodivergent, we really should do better.

Service

Workplace architecture

In the last few years, we at Arup have found ways to digitally map experiential aspects of buildings and environments, and then use the computational power of machine learning to envision different design solutions. Our team believes that formerly intangible considerations can actually be measured, unified, analysed and understood – using digital to better define our real-world, human needs.

Listen - what do we mean by neurodiversity?

Global Accessibility and Inclusion Lead, Mei-Yee Oram discusses what we mean by neurodiversity.

Play

Transcript

What do we mean by neurodiversity? (audio transcript)

When we talk about neurodiversity or neurodivergence, we’re simply recognising the natural diversity in human brains and the different ways we experience the world.

Today, we recognise that people can process information, learn and behave in atypical ways – for example autism, ADHD or dyslexia. For other people, with “neurodegenerative” conditions, their sensory processing capabilities may have changed over time from causes like dementia or Parkinson’s disease.

If we want workplaces to be neurodiversity-friendly and inclusive –  we need to learn from this range of experiences, and rethink the design of the modern workplace: That means the finishes and materials, and the nature of the products and fittings within the environment. It means the lighting, layout and acoustics. It also means thinking about the spatial design, and how different facilities are connected and located.

Then there are the ways a space will encourage connection between people, and with digital platforms, and the wider building itself. At Arup we approach these deeply human questions from multiple angles. We talk to users, to understand their experiences and challenges. And we’re taking advantage of new technology’s ability to model these nuanced and complex experiences. This approach shows we can design for neurodiversity and develop workplaces that work for everyone.

In the context of neurodiversity, this means analysing the day-to-day experience of sound, movement, physical layout and other factors – to help workplace designers bring neurodiversity into their methodology. This way we’re able to define and design a workplace that will be more appealing to employees, drive retention, creativity, and productivity.

Carrying out the analysis

Our approach to the modern workplace is informed by two core principles: one, we need to better understand and respond to people’s needs. Two, we should use data more productively, to shape what we know about an organisation’s real day-to-day needs as a productive and inclusive working environment. Given the complexity of neurodiversity how do we learn from these experiences, and let those insights reshape workplace design?

Tool

MassMotion

A key enabler is Arup’s in-house developed MassMotion agent-based modelling software. We have recently upgraded this platform to include a ‘social forces’ model that accurately predicts how people navigate through a space. We’ve developed a way to account for a variety of human experiences within a predictive model, drawing on personal testimony, sensors and data – this is a significant step forward for inclusive design.

Using machine learning to analyse this data reveals deeper insights into employee interactions within their work environment. This data enables us to develop predictive models customised to each client’s workplace, allowing us to simulate and evaluate changes and improvements. We believe this is the best way to bring neurodivergent users into the design process and deliver an experience that truly works well for everyone.

Listen - How can machine learning help us design better workplaces?

Architect and Design Computation specialist, Conor Black discusses how machine learning can help design the modern workplace.

Play

Transcript

How can machine learning help us design better workplaces? (audio transcript)

So you might reasonably ask, how can a computer understand the human experience of a modern workplace? This is where machine learning really demonstrates its potential. Machine Learning has the ability to process vast amounts of information– and can identify subtle patterns in large quantities of data – something humans cannot achieve unaided. We use a technique known as supervised learning – so, we train our model on many different aspects of a workplace, such as the daylight, proximity to amenities, layout of the space, and the preferences of neurodiverse employees. Then, using our own MassMotion human behaviour simulations, we run hundreds of analyses to build a picture of how different workplace design solutions will shape people’s experience. This then becomes our dataset – and it can predict how people will move around in a particular setting.

Traditional, more limited simulations used to take hours. Our machine learning analysis happens in milliseconds. The result is a new, finer grained level of behavioural insight, one that can help owners, designers, and users understand what kind of workplace they will experience every day. We believe this machine learning based approach can achieve even more though. We plan to add additional dimensions to the model’s understanding – including a space’s acoustic experience – in future. As we learn more about the experience of neurodiversity, we should be able to refine our models and offer a better design solution to meet the needs of everyone.

Many businesses have invested in smart building technologies, that promise to enable a more user-driven workplace, although this promise isn’t always born out in the real world. We believe our approach to neurodiverse workplace design demonstrates a powerful use case for smart building technology though – helping to drive future benchmarking and consistency across client estates, ensuring organisations can truly offer a great working experience to all their people.

A high angle view of people in a library.

“Nothing about us, without us”

We’ve long championed inclusive design – not just because it’s a sound ethical priority, but as a route to the most effective and valuable form of workplace design. It’s vital in this context, to let design be informed by actual human experience and testimony from those who see the world differently –  the motto “nothing about us without us” long used by the disability rights movement - is an essential principle for successful, inclusive design.

We believe that this approach is simply building out a new capability in human-led design, using machine learning’s power at analytics to better understand the complexity of human experience. It’s a cost-effective way of making a better long-term investment in the working environment that any organisation will depend on.