
REDD’s Space AI predicts the future. The next stage of the revolution
Which office will be rented the fastest and what to do to reduce the time when the spaces are empty? These questions will be answered by the new tool of REDD. The largest Polish start-up proptech has just completed the next stage of research work on the Space AI project. An algorithm developed by its authors based on millions of collected data will help property owners reduce the pandemic effect of “empty offices”. The company announces that later this year, real estate owners using the REDD online platform will be able to instantly verify the forecasted time of renting the space in their office buildings. The system will also suggest how to change the offer to speed up this time.
The past brings the future
In addition to over a dozen billion data on the Polish commercial real estate market, REDD will soon offer the possibility of extremely accurate forecasting of office renting. How? The future cannot be directly predicted, but very accurate models can be built based on… the past.
– In recent months, we have been working intensively on the Space AI algorithm. After many attempts, our machine learning model achieved an accuracy of over 75%, which means that in the set of historical transactions on the office market, the time needed to rent these spaces was accurately predicted in the vast majority of cases. This is a significant breakthrough, and maintaining such a level of accuracy in the future may be the beginning of a real revolution in the preparation of vacant space for tenants, developing negotiation strategies and broadly – managing the accounting liquidity of real estate – says Tomasz Ogrodzki, creator and CEO of REDD.
Space AI from REDD are models based on machine learning. As a result, a set of information (the so-called features) has been created, based on which the model predicts the time of renting offices. It is a very effective mixture of knowledge in the field of machine learning, statistics in the REDD database, and expert knowledge of the commercial real estate market.
Nearly a year to rent
The data from a year ago were not optimistic for the market. Office buildings in the middle of the “COVID era” were empty for almost a year (296 days according to REDD data) before another tenant arrived. Space AI is to change this, because apart from forecasting rental, it suggests which elements to optimise so that a given space is rented faster. The creators of the tool already confirm that the greatest potential lies in the highest class of commercial real estate, although the tool forecasts equally effectively in all classes of buildings.
The models clearly show that office rental is not governed by one universal law. Depending on the property class or its location, completely different parameters have a deterministic impact on the accuracy of our algorithms. Therefore, at the moment we are focusing on maximising the effectiveness of Warsaw A-class office buildings. In subsequent versions, we will expand to other markets and other asset classes – explains Mikołaj Synowiec, Machine Learning Engineer, REDD AI from REDD.
Space AI is another tool of REDD based on artificial intelligence that is revolutionising the real estate market. The REDD Group has also been working on a more holistic project for some time – REDD AI, which could soon revolutionise capital investments on the market.
– The AI market is valued at over USD 300 billion with the prospect of growth to USD 1 trillion in 2029. This value is the result of the wide use of artificial intelligence, or more precisely machine learning, in financial sectors, including banking, insurance and stock exchange. The algorithms determining, for example, granting of a mortgage, the amount of insurance premiums or trading machines are not science fiction, but standard business tools. Our goal is to implement this level of innovation in the commercial real estate market. How? REDD AI will allow investment funds to make strategic decisions faster and better – says Tomasz Ogrodzki, CEO of REDD Group.