An unforgettable journey in AI

A thought experiment into the potential of new technologies in real estate

In a 2020 research publication from AI21 labs, the average cost to train AI ranges from $2.5K to $1.6 Mio. for the average BERT* NLP**. *BERT is a deep learning model that is given a wide variety of NLPs or Natural Language Processing Techniques. **Natural Language Processing Techniques are processes with which AI extracts information from a natural language. An example of this would be automating a call centre and allowing AI to analyse the natural language of customer reviews.

Without AI, disaster can strike at any time. For the real estate industry, this could mean building the second largest building in the world only to find that half of the building is vacant with no apparent renters willing to move in.

This actually happened in 2018 after completing the Shanghai Tower that was in planning for more than 25 years. The skyscraper measures at over 600 metres with 128 floors and is an architectural
masterpiece.

The original goal was to capitalise on Shanghai’s growing economy and promote investors who wanted to move operations to China to jump on the developing economy. Unfortunately, the architects, real estate developers and experts all had one thing in common that held them back from seeing the red flags: they’re human.

As human beings, they most likely measured and mapped information for Shanghai Tower beforehand during the planning period to make sure nothing went amiss. However, the sad reality of being human is that a majority of indicators that can predict the future success of a property aren’t traditionally tracked according to a McKinsey study.

Seemingly abstract data get in the way of effective decision-making, and this is where AI can help. Machines can process data and find connections that are not as obvious to a human mind. On the other hand, humans might only be looking at a single point or two based on prior experience.

sander-crombach-On_ayJuJtgE-unsplash (1)

Beginning of the Journey: expanding intelligence vs strong AI

AI wasn’t readily available when the Shanghai Tower was built which lead to a miscalculation in demand that cost investors billions. For this article, let’s take a fictional tower project similar to Shanghai Tower, but provide our tower-building team with various AI to accomplish their goals.

Let’s assume that there has been a significant breakthrough in AI technology. In our theoretical world, companies regularly use AI to make decisions about how large a building should be, where the demand would be highest for the type of project and how the market will change in the coming years.

To better classify the type of AI that we’ll use for our thought experiment, we’ll use expanding intelligence. Expanding intelligence is the gathering, analysing, and presenting of information that could impact an industry. This is different from strong AI, which is regularly depicted in movies like The Matrix or the Terminator series and video games like Detroit: Become Human.

Thankfully, expanding intelligence won’t be taking over the world anytime soon as it is comprised of mostly natural language learning software that helps predict texts in things such as Gmail or Google Translate.

So, the question remains, what could our theoretical developers use expanding intelligence for to make their new tower not only profitable but beneficial to the community?

To start, our fictional real estate professionals would take their plans for the tower to AI-based companies like the German start-up EVANA. They have been training real estate-based AIs to help investors manage their assets while staying up-to-date with relevant market trends.

Stakeholders for our tower would work with AI-professionals to collect relevant data on the project to make edits on where the building’s location should be, how many floors it should have and what clients it might expect. If the height is the primary concern, an AI might suggest a city with a lower risk or an area within a bigger city that has proven to grow quickly.

Our fictional team of developers and real estate planners would work closely with data scientists and data engineers to reach into the data and pull out the information that backs intuitive decision-making.

This type of modern AI has a lower risk and higher satisfaction rates for large real estate projects among community members. Not only can AI predict profitability, but it can help place the right types of commercial or residential mix into these new buildings that have the best impact for both stakeholders and community members.

As our project continues, there will be many challenges the team encounters where expanding intelligence can help. Let’s focus on four that most real estate professionals encounter:

     1. Sourcing Leads
     2. Analysing Numbers and Risks
     3. Automating Contract Analysis
     4. Predict Market Trend

1. Sourcing Leads

Sourcing leads is one of the many daunting tasks a real estate professional has to overcome. Finding, qualifying and contacting leads can take hundreds of hours, and for our small team building the
tower, it isn’t realistic with their current capacity.

In our case, the team decides to look at AI solutions to lead sourcing and finds that many expanding intelligence algorithms can track the leads that are currently converting. Previous data helps predict future conversions from watching a real estate professional do their job successfully.

Our team scouts for the best AI to do the job and finds a company similar to Prospex where AI learns how to source leads for specific industries. Our fictional AI tool would sort through publicly accessible data to find investors willing to work with our team while building the tower. A platform might exist in the future for stakeholders to enter their information and be paired with projects, reducing the need for cross-platform data collection.

Our team can now spend more time communicating with stakeholders to finalise the project’s planning through a lead sourcing AI. Hundreds of hours of stress and uncertainty have also been saved, leading to a
stronger and science-based product.

2. Analysing Numbers and Risks

Now that our tower has an excellent start to lead sourcing, the team can spend their time on risk analysis. If the fictional tower takes as much time as the Shanghai Tower did, then there is a good chance the city we are building the tower in could change pretty radically.

For example, suppose we planned to invest in office spaces in the 2019 planning. It’s safe to say that unless
we change our plans that would still fit our architectural style, we might not be able to find businesses to rent the space as demand for offices goes down.

The same problem goes for retail. Still, what about the trends that we can’t see yet? This is where computers can help. There might be a growing push from the depths of data that tell us how the travel industry is changing and that new rooms are needed to accommodate this ever-growing market. Although it’s only the beginning of this emerging market, by the time the fictional building is finished, it would be in the best position to take advantage of this trend.

Analysing diverse data points in real estate is the most challenging task for our human investors as it is impossible to keep track of millions of data points. According to McKinsey, around 60% of market value indicators are not traditionally followed. This blindspot poses a considerable risk for our fictional team, so finding the right AI solution to monitor the data and provide data to make decisions is the best solution.

Expanding intelligence tracks these values automatically and creates detailed analysis reports based on an asset’s risk and future profitability. An example of this is the start-up Skyline which tracks over 50 years of real estate sales information to evaluate risk and profitability in the current market.

By automating this process, our team can make data-driven decisions that create a better strategy for real estate investors and reduce risk while increasing the potential growth.

3. Automating Contract Analysis

The fictional tower-building team now has a great tech set-up and enough information to begin working on the project confidently. As our team builds the tower, they find willing renters wanting to sign up before anyone else to make sure they have the space filled on the project’s launch.

Although it is possible to draft contracts for each type of asset be it commercial or residential, they decide to use a popular AI that drafts contracts that only require minimal edits. Normally, drafting contracts, analysing them and making sure they fit both parties’ needs is a painstaking effort.

With AI, they can fill in the details of the contract they would like, and the program writes contracts that fit their needs. After a short analysis, they can make small edits which the AI will keep track of for future use.

This AI is already common in today’s society with companies like Evisort. They use expanding intelligence to generate contacts automatically and learn from previous contracts.

4. Predict Market Trends

The tower is finished, and our team of real estate developers, architects, investors and other stakeholders are ready to tackle the next major project. They already have a market demand AI (also referred to as Demand Intelligence) that allows them to scan the market for demand gaps which are in need of new properties.

This AI predicts changes in the market and could have informed the Shanghai Tower creators of potential risks with market demand and growth. Suppose the investors had used an AI to predict the market demand for such a large building. In that case, they may have found that approximately half of the building would be left vacant after completion.

Our fictional team would depend on an AI-tool similar to that of today’s PredictHQ. They have developed demand intelligence that learns the client’s market to predict demand for products. Investors and business professionals can also use these algorithms to predict changes in the supply chain as well and have the power to increase overall global efficiency.

The future of AI in real estate

As expanding intelligence advances and becomes available to teams and investors, they can help guide the market through times of flux. AI can support dynamic decision-making in companies to put them ahead of the curves to come. Markets will benefit from higher-quality investment and products catered to their demand.

The real estate world can expect to develop new understandings in urban planning, building structures and community building through strategic developments from teams like our fictional one. Quality of life may also improve as AI discovers which neighbourhood make-up creates the ideal conditions for clients and biases are eliminated through the systematic editing of intelligence features.

he era of expanding intelligence will guide humans to produce better products for society and create long- term global benefits.

James Gibson

"We inspire the future of travel through a business model that works for new generations of travellers. By combining technology with quality design, we've been able to change what guests and investors expect from accommodation"

Join us by becoming a partner!