AI Design: Artificial Intelligence needs humans to be smart

Our ‘Future of work’ expert, Saimah Heron, reminds us not to get carried away with AI until we know it can pose the right questions.

Warming up your house before you get in can now be done at the touch of a button thanks to our increasing move to connected homes.

Smart homes are ones where technology is used to connect everyday objects and promote a frictionless life, and we’re buying into it – already one in four UK consumers owns at least one smart home device.

Using AI in real estate

This doesn’t stop at home. Artificial Intelligence (AI) is now being used to transform how we build and occupy real estate.

Challenger real estate company WeWork has embedded technology within its operations to such an extent that its £17bn valuation makes it look more like a technology than property company. It acquired Business Information Modelling (BIM) company Case in 2015, and has been using their technology to optimise office layouts and overall workplace design.

Taking the concept even further, Autodesk’s Project Discover surveyed 300 employees before generating a workplace design that accommodated individuals’ preferences, such as lighting and access to high traffic areas, enabling a seemingly tailored environment that suited their tastes.

Benefits of AI in design

Using BIM can allow for a rapid, iterative approach that is highly collaborative. The rich information modelled helps real estate companies sequence developments and visualise what a project will look like.

But is this approach to generative design focusing on the right questions?

Challenges when using AI

Perpetuating inefficiencies

Research has shown that AI can learn the same biases as humans. When it comes to building design, there is a risk that generative design replicates rather than resolves inefficiencies and quirks. Having robust reviews where design can be challenged are important to avoid perpetuating these flaws.

Determining the inputs

Can an individual predict how warm they want their working environment to be for the duration of a full day on that morning, let alone months or years in advance? Since the AI system is created by individuals, the inputs selected are chosen according to perceived utility.


Tailoring a space to a fixed group of individuals mean that as people join and leave, the original design inputs become irrelevant.

The danger is that we create systems that ask simple questions we can answer, whilst side stepping the complexity that would make a real difference to our working environments.

A different approach

Reflecting and responding to complexity is fundamental to how businesses should approach workplace design.

Business strategy

Avoiding a one-size-fits-all solution is best achieved by developing real estate solutions that fit with an organisation’s business strategy. Maybe it matters that the sales team is next to the marketing team, or maybe the bigger priority is how to embed health and safety principles across the whole working space. Priorities will be different for every organisation.

Spending time understanding how a company’s business strategy can be enabled by real estate is the first critical step before using AI to design it.

Fewer, better choices

Just as Apple narrows down its iPhone buying options to one choice between colours and another between sizes, rather than presenting endless customisable configurations, good design should focus on the choices that matter. This means investment in understanding what users should take for granted – good acoustics or a variety of work settings – before asking the right questions.

AI can be a brilliant tool but its outputs are only as helpful as the carefully designed inputs allow. Investing in the upfront and interpretative side of design can enable companies to create exciting spaces that reflect who they are as an organisation, and improve their overall business effectiveness.

If you’d like to discuss how planning for the future of work can support your business objectives, please contact me: hello@baxendale.co.uk.

Read the latest from our blog