Predicting the future – what could help us spot the next COVID-19 level shift?

This opinion piece was written by Arik Kogan, VP of Finance & Investment Solutions at MRI Software, and was featured in the EG UK Cities Investor Guide, published 3, June 2021. Read more here.

Machine learning and AI is going to be an extremely important part of our strategy at MRI Software. However, it’s not going to be a machine that anticipates the next major, COVID-level upheaval. Instead, I see software’s role as smoothing the path and freeing up headspace for analysts to apply human creativity and generate these kinds of visionary insights.

Machine learning provides us with more accurate data

A major priority for data accuracy is to reduce the need for human intervention and to ensure that data is complete, accurate, mapped appropriately and flowing into the right parts of the ecosystem. Machine learning is a big part of that – it can automatically gather, interpret, format and consolidate data much more quickly and accurately than a human could ever manage.

No matter how incredible our machine learning capabilities are across our solutions, there’ll always be exceptions requiring human intervention. Over time, though, I see an evolution of the processes we’re already using to reduce the need for human input in data gathering and mapping, so that analysts spend less and less time gathering and cleaning data, and more time using it.

Predictive analytics is the future of strategic planning

Predictive analytics is really the next big leap that we want to take.

Each company using solutions for planning and forecasting is doing things slightly differently, but there are also always things that remain consistent and repeatable regardless of circumstances – in areas of consistency like this, a lot of planning and analysis could be replaced by a predictive analytical engine powered by artificial intelligence.

For example – we can generally determine the cyclical aspects of the real estate market and their effects on things like interest rates, valuations, and transaction volume. There are aspects of these areas that could be programmed into AI, so a human being doesn’t need to go into a solution like MRI Investment Modelling to incorporate them manually – instead, the system can automatically build them into the plan.

So, for example, if you want to perform a capital deployment scenario and identify the best use of those funds, the system could use this information to build a strategy for you based on baked-in assumptions about target asset classes and regions based on market data and trends, and could make recommendations to reinvest in specific assets.

Could an AI have predicted COVID?

While predictive analytics could make successful recommendations in scenarios like this, there are elements where human factors make that impossible. For example, there’s no way an AI would have predicted the way COVID affected core assets.

An AI’s assumptions on core assets would be that they’re fairly stable, but COVID caused a global 30% drop in the value of core assets – offices, shopping centres and multi-family residential buildings – and corresponding jumps in areas like industrial buildings, warehousing and data centres. I don’t think you could design an AI, at least in the next 5 years, to predict that type of thing.

I don’t expect a computer to be flagging the next major COVID-level industry event. However, what AI can do is remove more and more of the “busy work” associated with data processing and modelling, leaving analysts free to model and test a much wider range of scenarios than they currently have time to, and to produce scenarios quickly to keep pace with board-level decision making during volatile times.

Laying a smooth path for visionary insights

At MRI, we know that while machine learning and AI are extremely exciting technologies, human creativity is the true differentiator for our clients.

To me, the real source of potential for real estate investment firms is the visionary insights their human teams come up with. The smoother and more intuitive our software processes can become, the more headspace we free up to allow talented analysts and decision makers to get creative with data and strategies.