Should you believe the hype around artificial intelligence (AI) for the real estate industry? In a word, yes. AI can add immense value for real estate organisations, because they generate large volumes of data and rely on it to make decisions.
Applying AI to PropTech will be the biggest change you never see. You won’t (or at least shouldn’t) be investing in AI for AI’s sake – but you should understand how it is improving PropTech to drive greater efficiency, better predict outcomes, and drive better business results.
When people hear the term AI, they often envision machines taking over and doing their jobs for them. But in reality, AI needs to work hand in hand with humans. It makes our jobs easier by shortening the decision-making cycle.
– Vijay Anand, VP, Artificial Intelligence
Despite the fascination and trepidation with auto-generative AI chatbots, AI is not a shiny new object. Let’s go back in time to the late 1980s, when a German chemical company called BASF launched a series of advertisements with the tag line, “At BASF, we don’t make a lot of the products you buy, we make a lot of the products you buy better.”
How does AI intersect with PropTech?
In its simplest form, AI combines computer science and robust datasets to enable problem-solving where we leverage exponential increases in computing power to tackle exponential growth in available data. AI is fueling the next iteration of PropTech, enabling us to take advantage of the data that we have been gathering and processing for the past 50 years and use it to drive efficiency, decisions, and service.
AI won’t be the PropTech you use, AI will make the PropTech you use better.
– Brian Zrimsek, Industry Principal
What are the specific activities and functions where AI can help? There are a number of areas where AI and PropTech intersect, many of which are already available today.
Using AI for predictions and recommendations
Using AI to make predictions is a task perfectly suited for the combination of robust computing power and vast amount of data available today. Understanding how key metrics relate to business outcomes can help drive more positive results.
– Vijay Anand, VP, Artificial Intelligence
Many of the following examples leverage historical data to predict or influence human behaviors and provide recommendations designed to drive better outcomes.
AI can help real estate firms make better predictions in several areas:
Applicant screening
Better understand both an applicant’s ability to pay rent along with their propensity to pay rent. This gives you a clear picture of the risk an applicant represents as input to the applicant approval process. AI can also be used to verify an applicant’s income, interrogating banking deposits over time to accurately represent income.
Pricing
The array of supply and demand variables utilised to recommend market pricing is substantial. Best practice is to frequently tune pricing recommendations to current conditions, which is similar to longest established practices in the hospitality industry.
Propensity to renew
Similar to market pricing, understanding the relationship between the lease renewal increase offered and the likely renewal outcome also benefits from using historical outcomes to generate renewal offers that are in line with business objectives.
Predictive maintenance
Understanding mean time to failure as well as current operating parameters for key assets allows for preemptive action before failure and potentially larger issues occur. These work items can be injected into backlogs to optimise resource deployment. AI can also determine if a replacement should be considered in place of a repair.
Usage variances
Collecting and analysing time series data from IoT (Internet of Things) devices and other sensors provides a great baseline for the detection of operational anomalies. Abnormal spikes in energy usage, water flow, facility access at odd hours, changes in vibration, and other measurable parameters can all be leveraged to alert maintenance and facilities staff that attention is needed if something is operating outside of the norm.
Arrears forecasting
Understanding typical payment patterns and correlations to other available data points helps predict future risks in existing commercial relationships and allows for more targeted engagement, especially where risk is growing.
Maintenance routings
The workload for a maintenance technician can be considered in terms of criticality, activity size, and location to generate a recommended sequence of activities and optimise resources while achieving customer service goals.
Processing, prevention & reaction
AI can also be used to render an action as opposed to simply making a recommendation. Data can be utilised
to automatically advance a business process, detect inappropriate content, or prevent specific issues based on observable deviations from standard data patterns.
Process advancement
In many cases, AI can be trained to follow a rule set and take specific actions based on specific conditions without a need for human intervention. This could manifest in the auto coding of invoices, streamlining the bank reconciliation process, replying to customer inquiries, or other repetitive tasks that are necessary but are more mundane than value-adding.
Image comparison
One critical example is the use of image comparison to prevent fraud. Using AI to compare an image from a government issued ID to a selfie photo submitted by the same party can identify a match or uncover someone trying to misrepresent their identity.
Energy consumption
Matching the environmental conditions in the building to the ongoing usage patterns offers an opportunity to be more environmentally-friendly while providing better conditions for occupants. Managing heating, cooling, lighting, and even elevators to best match ongoing usage patterns can reduce overall energy consumption, especially when hybrid working is becoming the new norm.
Content moderation
Interrogating images for specific content is a common use for AI. Training a system to understand what is acceptable and what is not allows for efficient management of large volumes of publicly sourced submissions. Ensuring that people are not included in property images and disallowing explicit images are two such examples of content moderation that are directly applicable to PropTech.
Communications & engagement
Leveraging AI to enhance customer service and engagement continues to be an area of focus as organisations seek to drive traffic with marketing, convert that traffic to leads, and then nurture those leads through the lead-to-lease process.
Chatbots
Using natural language processing to engage with prospects offers the ability to answer questions in real time and drive the prospect to schedule an appointment or to complete an application. By some accounts, chatbots can cover over 90% of the questions a prospect may have, freeing up property staff to focus on more valuable interactions. This same technology can also be used to ensure that the needs and questions of current occupants are addressed.
Prospect engagement
Between the time when someone becomes a prospect to when they convert to a resident, there are a number of activities that need to be managed to maintain engagement and move the prospect through the funnel. In place of waiting for a human to send the next communication, an AI-powered leasing assistant can continue to engage the prospect by providing timely communications, especially when the answers to questions are well understood and repeatable. Again, we are using AI to free up the human to focus on more valuable tasks without sacrificing customer service.
Insights from analog inputs
In most of the examples above, AI is using well-structured data to drive better outcomes. Data in less structured formats, including analog, can also be harnessed by AI as noted in the following examples.
Extract data from documents
Taking data from an unstructured format in a document into a structured, digital form unlocks significant business value because what once required manual searching can now be extracted and parsed digitally for faster reference and use in other applications. This type of AI can be applied both to highly bespoke documents like leases and contracts and also to more formatted documents like tax forms and certificates of insurance.
Understand pedestrian patterns and demographic
Using AI to understand pedestrian traffic patterns, including paths taken and dwell time, can unlock a vast array of valuable information for retail tenants and property owners. Add to that the ability to attribute demographics to the observed traffic and you vastly increase the value of the data.
As we look toward the future, you can expect AI to continue to blend into PropTech – it won’t be the solution you buy, but it will make the solution you buy better.