Big data, artificial intelligence, proptech and digital transformation get a lot of attention and rightly so. Data from Google shows that all of these terms have experienced significant growth over the past five years in particular.
Using just the lens of the property industry, it’s clear that many organisations in property are sitting on vast, ever-growing, quantities of data. But it’s not just about data.
In the next few years, data will become increasingly available. Leaders in the real estate industry will need to evolve quickly to unlock meaningful insights and thereby provide a better user experience for tenants, investors and customers – or risk being left behind.
At Investa, we are actively investing in data and decision intelligence. We believe that it is first about asking the right questions, then having the right people, data and technology to find the answers. Investa believes that organisations will derive a real competitive advantage by building decision intelligence capability, and that the opportunity is “now.”
Forward-thinking property companies will progressively be able to see and realise value that the general property market cannot see. As our models and data sets become more sophisticated over time, we will see increasing levels of return on the investment in this capability.
Part 1 of this article defines AI, Machine Learning (ML) and Deep Learning (DL) and looks back at its history, how we got here and how we can benefit.
Part 2 looks at six force multipliers that businesses use to drive a ROI from data and AI.
Part 3 focuses on some of the challenges and risks that come with AI.
Part 4 outlines some specific real estate use cases of AI.
2. What is Artificial Intelligence (AI)?
Artificial intelligence refers to the wide-ranging simulation of human intelligence by machines.
The goal of AI is to solve the kind of problems or perform the types of tasks that are usually completed by humans, with our natural intelligence.
There are two main types of AI: The first is “Weak AI” or Artificial Narrow Intelligence; and the second is “Strong AI” or Artificial General Intelligence.
AI will typically demonstrate some of the following characteristics: planning, learning, reasoning, problem solving, perception, motion, manipulation and to a lesser extent, social intelligence and creativity.
According to Artificial Intelligence: A Modern Approach, by Stuart Russell and Peter Norvig, AI can be defined by four fundamental approaches:
1. Thinking humanly
2. Thinking rationally
3. Acting humanly
4. Acting rationally
The first two approaches centre around thought processes and reasoning, the latter two focus on behaviour.
3. What is Machine Learning (ML)?
Understanding the difference between artificial intelligence, machine learning, and deep learning can be confusing.
Venture capitalist Frank Chen made this distinction:
"AI is a set of algorithms and intelligence to try to mimic human intelligence. Machine learning is one of them, and deep learning is one of those machine learning techniques."
ML feeds a machine (computer) data and uses statistical techniques to help it learn how to progressively get better at a specific task. Machine learning consists of both supervised learning and unsupervised learning.
4. What is Deep Learning (DL)?
Deep Learning (DL) is a type of machine learning that runs data through brain-inspired neural network architecture. The neural networks allow the machine to go deep in its knowledge-base, making connections and weighting data.
5. What are the different types of artificial intelligence?
Artificial Intelligence falls typically under two categories: Narrow (or weak) AI and artificial general intelligence (AGI) or strong AI.
a) Narrow (or weak) AI
Narrow or weak AI operates within certain limitations and is a simulation of human intelligence. It's often focussed on executing a single (or narrow) task really well.
Much of narrow AI is powered by machine learning and deep learning. This AI can often outperform humans in a specific niche task — especially if the task relies on historical knowledge and data.
Some examples of narrow AI include technology that is broadly used: Google, Siri, Alexa, self-driving cars, chatbots, translators, social network recommendations. The most famous example here is IBM's Watson.
At Investa, we have been using AI/ML in our business for years. For example, we use cybersecurity AI software to detect potential threats, and search engine AI across our internal files and emails.
In addition to these baseline use cases, we are focused on narrow AI use cases that make an impact on our business, customers, investors and employees.
For example, identifying new development sites to acquire, or what tenants are the best fit to occupy our buildings. This application of narrow AI allows us to augment what we do as humans and remove some of the repetitive tasks in our business.
b) Artificial General Intelligence (AGI) or strong AI
Artificial general intelligence is the creation of a machine with human-like intelligence that can be applied to any task. It seeks to replicate the cognitive abilities of the human brain.
When presented with an unfamiliar task, a strong AI system should be able to apply knowledge from another domain or task to find a solution autonomously.
Strong AI systems tend to be more complicated, and is often the subject of science fiction. We are not yet deploying AGI into our businesses, but it is definitely on our radar for the future.
6. What is the history of artificial intelligence?
Artificial intelligence had its start in antiquity by mathematicians and Greek philosophers. But when we think of artificial intelligence in modern-day terms, its history spans less than a century. Here are some of the key milestones in AI.
1943 – Warren McCullough and Walter Pitts publish A Logical Calculus of Ideas Immanent in Nervous Activity. The paper proposed the first mathematic model for building an artificial neural network.
1952 – Arthur Samuel wrote the first game-playing program for checkers with sufficient skill to challenge a human player.
1958 – John McCarthy develops the AI programming language Lisp and publishes the paper Programs with Common Sense. The paper proposed a complete AI system with the ability to learn from experience as effectively as humans do.
1981 – Digital Equipment Corporation develops R1 (also known as XCON), the first successful commercial expert system. Designed to configure orders for new computer systems, R1 kicks off an investment boom in expert systems that will last for much of the decade, effectively ending the first “AI Winter”/ R1 saves the company $40 million a year by 1986.
1997 – IBM’s Deep Blue chess machine beats world chess champion Gary Kasparov.
2002 – iRobot creates Roomba, the first commercially successful robot for the home. It autonomously vacuums the floor while navigating and avoiding obstacles.
2008 – Google launches an app on the new Apple iphone with speech recognition.
2014 – Google makes the first self-driving car to pass a state driving test.
2016 – Google DeepMind’s AlphaGo defeats world champion Go player Lee Sedol. The complexity of the
ancient Chinese game was seen as a significant hurdle to clear in AI.
2018 – Alibaba developed an artificial intelligence model that scored better than humans in a Stanford University reading and comprehension test, scoring 82.44 on a set of 100,000 questions.
7. Why is AI popular now?
Increased computer power has been a significant driver for AI, especially infrastructure speed, availability and scale. What used to be run in specialised labs with access to supercomputers, can now be deployed on the cloud at a fraction of the cost.
Thanks to critical mass and the awareness of natural language personal assistants, like Siri and Alexa, AI has also become mainstream. Large players are investing heavily in it and AI has fuelled an explosion of AI-based start-ups across industries such as healthcare, finance, insurance, real estate, commercial real estate and many more.
Big data is another driver for AI in recent times. The increase in data has created a strong demand for solutions that go beyond simple data analysis and instead promise deep, new insights and intelligence. Data is the fuel for AI, and as data sets become more accessible, higher quality and cost less, the barrier to entry for building effective AI models decreases.
As real estate is the largest asset class in the world, there is a large amount of data that we can use for our use cases at Investa. This data is becoming better, but is still disparate, incomplete, hard to access and difficult to structure.
Investa partners with data providers such as Empirical CRE to build robust data sets, but there is a large amount of work, creativity and knowledge required to hunt and structure the data we need.
“Investa understands the benefits that can be derived by leveraging both the intelligence in our data, as well as in our people, to ultimately drive business value.” Jonathan Callaghan, CEO, Investa
“Before Empirical CRE entered the market (commercial real estate), data providers focused on limited groups of properties using the basket approach, which in turn provided a limited picture of certain datasets without access to real-time or frequent captures of trends. Empirical CRE's objective is to track all properties regardless of size, geography, or type across all major metro areas to provide landlords, agencies, and their research teams a platform with standardised content to aid critical decision making in one of the highest value real estate markets in the world,” said Doug Curry, EmpIrical CRE.
As the data in real estate matures, increasing numbers of property companies will be able to access and derive value from these data sets, which will allow them to leverage AI/ML even further.
8. What are the benefits of AI?
Humans and AI work like a partnership.
AI excels at complexity and finding patterns in vast quantities of data – which would be impossible for a human to compute. Humans excel at creativity and ambiguity — using the insights from AI to support decision making and negotiation.
Investa has partnered with cutting-edge AI platform provider SparkBeyond, to accelerate and amplify the way in which the business collects, organises, analyses and draws insights from data.
"Using AI to generate data-driven insights to inform decision intelligence is a cultural shift for many organisations. In our experience, doing it well requires strong collaboration across the organisation to allow AI and humans to play to their strengths. It has been a pleasure to work closely with the Investa team to support them through this cultural shift and we're excited to see this translate into real business value.", Katherine Leong, Impact Strategist, SparkBeyond
The benefits of AI complement rather than replace humans. Here are some of the benefits of AI.
a) Reduced “human error”
Humans as we know aren't perfect and we make errors.
Using AI, computers can avoid (or reduce) making these "human errors," providing greater accuracy and precision. This has the effect of increasing productivity in the economy and driving growth.
b) Less risk
Reducing the risk to human life is one of the most significant advantages of AI.
Robots driven by AI can be used on Mars, to defuse a bomb and even for natural disasters.
From a business perspective, AI can remove risks by flagging anomalies like fraudulent transactions, it can monitor equipment and it can even suggest preventative maintenance.
Another benefit of AI is the sheer workload it can carry.
Where the average human can work for 8 hours a day, including some breaks, AI machines can be operational 24/7 and without any downtime.
For the property industry, there is a benefit for use cases that require constant monitoring — like optimising energy usage in a building for example.
d) Repetitive jobs
As part of life, humans perform many repetitive jobs. Think of manually inputting data into Excel spreadsheets, replying to emails or verifying documents.
With deep learning and machine learning, AI can “become smarter over time.” It can take care of repetitive tasks and increase business efficiency.
In the property industry, AI can be used to analyse tenant requests and anticipate what types of requests will be logged (and when). This allows organisations to better predict tenant issues and enables them to provide better, proactive customer service.
e) Digital assistants
In the realm of sales, marketing and customer service, AI can be used to augment human interaction.
For example, chatbots can be used on a website and social media to engage with users and pre-qualify them with questions that can better direct their enquiry within the business. Chatbots can also operate 24/7, allowing users in different time zones to chat with a business at a time that suits them.
At Investa, we are starting to experiment with digital concierge, which can provide service at our buildings outside of standard business hours..
f) Faster decision making
Thanks to the sheer amount of data points that AI can process and analyse, AI can enable faster decision making.
For example in September 2019, Investa partnered with AI-platform provider, SparkBeyond.
The purpose of the partnership was to accelerate and amplify how Investa collects, organises, analyses and draws insights from data, at a scale not previously possible.
AI enables Investa to harness previously untapped insights from hundreds of internal, external and open-source data points, enabling unprecedented decision intelligence across a range of core business areas including acquisitions, tenant attraction and retention, pre-emptive building operations, optimal deployment of capital and aligning investors with the right opportunity at scale.