Artificial Intelligence (AI) is a highly used term in proptech these days. Its implementation, its uses, and its decision-making processes remain a mystery to many, so I need to clarify a few things before starting.
It is important to understand that an automated process is not artificial intelligence. An automated process is a program based on a set of rules to complete a task and is reactive, whereas, AI is based on analytics, and for AI to work, it needs a lot of data. AI makes an automated process intelligent.
All of this talk about AI reminds me of the mid-2000s when I was starting my FEDGOV cyber operations company, ISHPI. While many of my competitors were operating in the information assurance (IA) space, my company was operating in a highly specialized cyber space within IA and was growing exponentially. To compete, these companies started to market themselves as full-scale cyber companies – cyber sold better! It did not take long for clients to realize these companies were making false claims about their capabilities.
I mentioned cyber because I see a similar thing happening with AI in proptech. Companies are claiming to use AI when they are using automated processes instead – AI sells better! There are companies like Restb that use AI, but look at the iBuying platforms like Opendoor, Zillow Offers, Offrs, Redfin, Realogy, and others that all use AI-backed algorithms in their pricing models. Where are they now? They are out of the market because their AI-backed algorithms hit a speed bump. For these AI-backed algorithms to work properly and efficiently, they need a lot of (good) data. COVID-19 disrupted that good data.
How do I know? One of the ventures I am involved in, launched into alpha in early March after eighteen months of AI, systems development, training, back-testing, and an alpha pre-launch test. Everything was working flawlessly. We were happy. Our investors were excited. Three days later, COVID-19 was announced as a global pandemic and our model had to be shutdown. Models did not have COVID-19 built into them, but as the pandemic passes and the data begins to normalize, AI will be ready for the next disruption (whenever that may be).
There is a lot to learn about AI and its uses in real estate. Using MLS data as an example, our trained models and AI discovered that Multidimensional Seasonality™ exists in these enriched datasets, which has made our best-in-class CMA solution even better. Unfortunately, some companies are using AI to develop shiny objects that do not solve any problems or create any efficiencies. Many of these companies are spending a lot of money without fully knowing the ROI. If it solves a problem and brings value to the client, then build it – they will buy it. We implemented AI to create a better solution for our product.
Not all AI has been affected by the pandemic.