Real Estate Research
How will Artificial Intelligence affect real estate?
While there is a lot of hype around AI and many applications still need to be refined, it is likely to have far-reaching consequences for real estate.
Whether its self-drive deliveries, chatbots, or warehouse robots, we foresee some far-reaching consequences for commercial real estate from the rise of AI.
Artificial Intelligence (AI) has been around for decades, but has only gone mainstream relatively recently. Its impact on a wide array of sectors is well-understood, but on commercial real estate perhaps less so. So, how might it influence retail, industrial and office markets?
The current state of AI
First, some background. AI, sometimes called machine learning, can be defined as the ability of computers to use data to find patterns and make predictions without relying on pre-programmed rules and instructions.
After decades in the computer lab and several false starts, “narrow” AI, which can perform a specific task, is being applied in a number of different fields. These include online marketing, augmented and virtual reality, and “expert systems”, which can simulate the decision-making ability of an expert in a particular field.
Artificial “general” intelligence, which is the ability to switch from one intellectual task to another, like a human, is still some way away.
The schematic diagram below breaks AI down into a number of separate applications. Those applications, which are already mainstream, are shown on the right, while those, which have not yet achieved mass adoption, are to the left, with the dashed line representing the inflection point.
AI: where is it mainstream, where is it not?
While all of these innovations are valuable, probably the real prize is computer vision. This will enable machines to recognise people, identify objects and anticipate how people and other machines are moving around them. Once computer vision is reliable it will be possible to build “smart” robots which can safely work along side people and a lot of AI technology is also being incorporated into autonomous vehicles.
What’s driving AI?
The progress of AI is being driven by two main factors.
The first is the exponential growth in data from connected devices such as mobile phones, tablets and the sensors, which are now built into industrial equipment, transport equipment and household appliances (i.e. the Internet of Things). The number of connected devices is expected to increase fourfold between 2016 and 2025 and the amount of data by tenfold.
The potential insights, which can be gained from data, mean that big companies and organisations, which have large volumes of data typically, have a headstart over small companies, although data quality also matters.
Second, the development of AI is being facilitated by advances in graphics chips, which mean that algorithms can process data in real time.
Predictions not explanations
However, while AI is a general purpose technology which will lead to improvements across a wide range of activities including banking, marketing, medical diagnosis, security, etc, it will not necessarily further our understanding.
For example, Walmart the US retailer has found a strong correlation between hurricane warnings and sales of strawberry pop-tarts, and analysis reveals that people who make a point of promising to repay when applying for a loan are, perversely, more likely to default. However, no one is sure what causes these relationships.
In short, AI is in the prediction business, not the explanation business3. That might not matter for groceries, but it could be problematic if doctors cannot fully explain an AI diagnosis to a patient, or if the algorithm, which a company uses to vet job applications, appears to discriminate against certain candidates.
Commercial real estate and AI
How will AI affect commercial real estate? At this stage, it is difficult to be certain, but there are some pointers.
In the retail sector, AI should improve the in-store experience for shoppers. The need to wait to pay for goods should disappear as object recognition cameras identify the items in peoples’ baskets and as facial recognition and mobile payments combine to take their money as they exit the store and e-mail a receipt. KFC has already experimented with a “Smile and Pay” terminal in China.
In addition, stores will have screens to give in-store directions, display goods in augmented reality, provide detailed product information and create customised products. For example, L’Oreal has developed a scanner, which can analyse someone’s skin tone and then dispense a personalised foundation.
In the office market, the combination of AI and data from the Internet of Things should enable occupiers to use space more efficiently and reduce energy consumption. While this could cut demand, we believe that the impact on head office, or front office space will largely be offset by the increase in informal working areas, as companies look to attract young, highly skilled staff, promote internal collaboration and improve productivity and well-being. More generally, the growth of AI and other technology (e.g. blockchain, cyber security, social media) will increase the demand for space in the tech districts of certain big cities (e.g. Amsterdam, Berlin, Boston, London, Munich, San Francisco, Seoul, Singapore and Stockholm) and university cities However, on the downside, AI will cut the demand for back office space, as chatbots become more proficient and start to take over a lot of the routine enquiries which are currently handled by customer services teams and call centres. Furthermore, expert systems are likely to replace people in functions such as invoicing, selecting job applicants, insurance claims, loan approvals, compliance and product design. AI is therefore likely to add to the existing pressure on administrative jobs and back office space from off-shoring, robotic process automation and blockchain.
Industrial real estate
To date, the industrial real estate has been the clear winner from technology, as the rapid growth in online retail has boosted demand for both big distribution warehouses and smaller, urban logistics units for “last mile” delivery.
Looking ahead, AI will enable robots to progress from just moving goods within warehouses to picking and packing and lead to the development of more sophisticated delivery drones which can decide what to do if no one is at home. However, AI could also start to undermine the demand for warehousing, as better information allows manufacturers and retailers to predict sales more accurately and turn stock over more quickly. Leroy Merlin, the French DIY retailer, cut its inventory by 8% following the introduction of AI system for re-ordering products5.
Finally, what impact will autonomous vehicles have on commercial real estate, once they become mainstream, probably some time after 2025?
One possibility is that people will simultaneously decide to stop owning a car and switch to hailing a car when required. In this scenario, cars would spend most of their time on the road and there would be a lot of surplus parking spaces which could be re-developed into something more valuable.
Another possible consequence is that more people will choose to commute to work by car rather than by public transport, because they can travel door-to-door without concentrating on the road. If so, then the rental premium which currently exists for shops and offices next to railway, or metro stations is likely to fall. Likewise in the logistics sector, driverless trucks could change the optimal location for big distribution warehouses, because lorries will no longer have to take a rest break every four and a half hours.
Conclusion: believe the hype?
While there is a lot of hype around AI and many applications still need to be refined, it is likely to have far-reaching consequences for real estate. On a positive note, AI should improve the in-store experience for shoppers, increase efficiency and staff productivity and give a further boost to the demand for office space in tech clusters. Conversely, autonomous vehicles are likely to change the optimal locations for shops, offices and warehouses and AI will accelerate the decline in back office space.
1.Artificial Intelligence: The Next Digital Frontier? McKinsey Global Institute. June 2017.
2.Data Age 2025, IDC / Seagate. April 2017.
3.“Everybody Lies”, Seth Stephens-Davidowitz. 2017
4.The Turing Test is when people are unsure if they are communicating with a computer, or a human.
5.The Economist, 31st March 2018.