Everything everywhere all at once: the RentLondonFlat view on AI

This article was originally prepared for a spotlight piece on RentLondonFlat.com Managing Director EuGin Song that appeared on 17 May 2022, published by the UK PropTech Association.

An apt movie title to describe the growing pervasiveness of artificial intelligence. The range of technologies and solutions available continues to grow at an exponential rate. Just take a look at Pantheon Lab’s Aidol.ai which has created a human presenter.

The girl in their video is not real.

Seriously impressive.

And yet, this is only one use case. There are so many other applications and use-cases for machine learning, and AI more generally. The number of models being created and applied to visual, audio, text, and tabular data sets mind-boggling and hard to keep pace with. But most critically, they are becoming easier and easier to put into production and deploy with each passing month.

When AirBnB and Uber started, they had to invest millions of dollars in building custom models on monolithic code bases programmed line-by-line. However, today we have cloud-based microservices able to rapidly deliver machine learning models, developed on huge data sets, by simply connecting our data rails to their APIs. Today, fast deployment, testing and scalability is available to any business anywhere in the world. It is a complete game changer and presents massive opportunities for businesses able to navigate this new frontier.

Meanwhile, the property industry is often criticised for being a laggard when it comes to adopting these new technologies, with an incumbent culture resistant to change. However, I think this is somewhat unfair, as it does not take into account the nature of the industry itself. The underlying data sets the property industry has to deal with is far more varied and complex than the FinTech or InsureTech industries.

For example, I used to build credit scorecards and business simulations for banks like NatWest and Barclays. For any data I needed to train a model, I would simply connect to the bank’s mainframe and download the required datasets. In those days, we had to analyse data line-by-line between code and spreadsheets. However, those kinds of data sets are not as common in the property industry. They have to be constructed and are spread between both private and public data silos. The data is far more fragmented and has huge variability in data quality. More time is actually spent on pulling those data rails together and cleaning them up rather than actually building and training custom AI models. Automation has to happen before AI can be truly useful.

The good news is, that change is happening, and it is disrupting the way the property industry works and delivers services. The pace may feel glacial in some quarters ,and the change imperceptible on a day-to-day basis. However, in the background, there are seismic changes happening. The intersection between IoT, blockchain and AI is truly exciting. The opportunities are endless. You interact with AI everyday on your phone without knowing it. AI will be in everything everywhere all at once.

You may not notice it, but you will definitely benefit from it as services will become more efficient and less costly. Middle men and administrative human labour in offices will become increasingly dis-intermediated, in the way the industrial revolution moved human labour from farms, to factories. AI will move labour from offices to leisure and home. That is what technology does. It is deflationary by nature and something we very much need now.