AI and ML driving innovation in real estate investment
"The ad said this pool was lagoon-like. There's nothing lagoon-like about it, except for the bugs." That's the sarcastic comment of two clients in an iconic scene from American Beauty right before Carolyn, the movie co-protagonist working as a real estate agent, has a meltdown.
Despite the typical cinematic overdramatisation, it's actually true that few industries in the world are driven by the "human factor", including interpersonal interactions between sellers and buyers, as much as real estate, and this seems to result in a certain conservatism and low propensity for change. Just a few years ago, indeed, this sector was ranked as the second least digitised by the Morgan Stanley Digitalisation Index.
However, things are evolving faster, as relying on technology to compete in a saturated, unpredictable market like real estate can certainly come in handy. Let's explore the opportunities that may be unlocked through digitalisation, especially by implementing artificial intelligence (AI) and machine learning (ML) in real estate.
What is AI in real estate?
AI in real estate encompasses a full spectrum of technologies deployed to spot the most profitable investment opportunities on the market, easily connect sellers and buyers, streamline property management, and minimise business risk. These generally include:
- Cognitive technologies such as NLP (natural language processing) and computer vision, combined with networks of IoT (Internet of Things) devices to collect information from textual or visual data sources.
- ML leveraging powerful self-learning algorithms that analyse such data, unveil the real estate market dynamics, and help make informed decisions.
- AI-powered software bots to automate and speed up a variety of marketing and administrative tasks
AI and ML use cases in real estate
How to harness the full potential of AI-related technologies in terms of data extraction, analytics, and automation? Here are three business functions of the real estate industry that are set to undergo a radical process of renewal thanks to artificial intelligence and machine learning.
1. Real estate appraisal
Settling a great real estate deal largely depends on an accurate estimate of a property's value and potential rental price, which was traditionally performed by relying on a combination of "gut", experience, and relatively primitive analytical methods. These could encompass variables such as square footage, location, asset's key attributes, and historical data regarding previous purchases.
Needless to say, such a limited investigation may not be enough, especially in a market where personal perceptions and preferences, along with broader trends and dynamics, can be rather mysterious and challenging to decipher. That's where AI and ML come into play, specifically in these two ways:
- Data collection: AI-based cognitive technologies can harvest a wider range of data types coming from a multitude of sources and therefore expand the list of parameters taken into account for our analyses. For example, computer vision is used to scan satellite imagery, property photos, and other visual content. NLP, on the other hand, can be deployed to peer into social media, online articles, government reports, competitors' ads on real estate platforms, and more.
- Data analysis: Machine learning algorithms are able to process the aforementioned datasets, identify a more comprehensive set of variables and the recurring patterns or relationships amongst them (including underlying dynamics which would be impossible to pinpoint for "mere humans"), and forecast how such metrics may impact the value of a real estate asset.
Regarding this increase in data sources and parameters considered, it's worth pointing out that the adoption of machine learning-based data analytics tools enabled realtors to keep an eye on so-called non-traditional data, which may power, based on McKinsey's estimates, 60% of our predictive capabilities. Non-traditional variables can complement standard information to provide a full picture encompassing property features, asset's listing status and legal information, market trends and financial data, socioeconomic metrics (such as quality of services, unemployment and crime rate in the neighborhood), and more.
2. Marketing optimisation
Without an efficient tool to spot potential deals and offer a communication channel between sellers and investors, the real estate market would end up unmoveable... like a real estate property. Once again, AI and ML offer a relevant contribution by streamlining several marketing-related functions:
- Targeted advertising: This well-known embodiment of machine learning in marketing, commonly adopted in ecommerce and many other sectors, can also be deployed in real estate. It involves segmenting customers according to their user data, browsing history, and activity on social media and online platforms, pinpointing recurring behavioural patterns to frame their interests, and targeting them with the most suitable real estate ads.
- Chatbots: Fuelled with AI and natural language processing, these tireless assistants provide potential customers with 24/7 assistance by answering questions in real time and offering relevant information on the most interesting real estate deals in a certain area.
- Marketing automation: Bots can also automate various marketing procedures, such as the compilation of mailing lists or even the publication of social media posts and other types of content depending on specific parameters to maximise their reach.
3. Property management
Investing in a real estate asset doesn't end with the mere purchase of a property. On the contrary, it requires ongoing administrative and maintenance efforts both to maximise profit in the long term and minimise the potentially risky implications of ownership. Here are some processes which can be better handled by implementing AI and ML in property management:
- Predictive maintenance: Machine learning-based systems, combined with IoT devices spread around properties, can gather data from sensors and cameras, identify deviations from the standard functioning patterns of the electrical system or any other infrastructure elements, and therefore report such anomalies to the management for timely interventions. A similar procedure can be performed to monitor and therefore optimise energy consumption.
- Fraud detection: A different but equally dangerous type of anomaly which can be spotted by ML algorithms concerns fraud attempts, typically identified by cross-chequing bank transfers, criminal records, and account data in search of inconsistencies. This represents an essential element in tenant onboarding and KYC procedures to mitigate business risk.
- Finance and accounting: Similarly, to marketing automation, bots combining RPA (robotic process automation) technologies and AI can easily perform a variety of time-consuming clerical tasks faster and more accurately than humans. These include lease contract processing, communication with residents, cost management and consumption-based invoicing, tax reporting, NAV calculations, etc.
From conservatism to a new tech momentum
Over the years, the real estate industry has been defined not only by the obvious pursuit of profit but also by a strong human and relational component that may have slowed down the digitisation process experienced by other sectors. So much so that, according to a 2020 report by the National Association of Realtors, 89% of American sellers relied on an actual real estate agent to sell their home and 41% of them still chose a realtor based on friends’ and relatives' recommendations.
On the other hand, several real estate platforms and companies such as Zillow, Redfin, HomeUnion, Opendoor, and Airbnb have already begun to integrate impactful AI-based features, leading by example and paving the way for wider adoption of artificial intelligence and machine learning in this field.