Decoding AI and Machine Learning: Key Insights from Our Webinar

Proptech Australia recently hosted an insightful webinar on the growing role of AI and machine learning in the property technology sector. 

The event was led by Proptech Australia President Kylie Davis, and featured an expert panel of  Damian Merchan of Sonar AI, Aonghus Stevens of Asseti, Leann Jones of Nimo Industries, and David Howell of Dynamic Methods.

The discussion explored four distinct ways AI is emerging in proptech, offering valuable takeaways for both tech providers and their clients. 

Here are some of the key insights shared during the session:

A Customer-First Approach

The panel emphasised that proptech solutions should be customer-first, not AI-first. 

This means building a solution around a clear user need rather than starting with a technology and trying to find an application for it. This approach ensures a strong, viable use case that provides genuine value to an internal process or to the end-customer.

Transparency is Key

The panel highlighted the importance of transparency, urging a "Clear Box, not a Black Box" philosophy. This means that both proptech providers and their clients should have a clear understanding of how the AI system works. 

Providers should be transparent about their data processing, security policies and the partnerships they have in place. Clients should feel empowered to ask clarifying questions about these protocols.

The Importance of Context and Data

A critical point of discussion was the role of data and context. The panel noted that a client's existing processes and systems are a valuable part of the data that an AI strategy must accommodate and often incorporate. 

It's crucial for clients to understand where their own data begins and ends, and how a proptech’s data ecosystem will enhance or improve it.

A particularly interesting concept raised was “context rot”. This refers to the gradual degradation of an AI model's performance as the real-world context it operates in changes over time. 

For example, an AI trained to predict property values based on market trends from a stable economy might perform poorly during a recession because the underlying economic context has shifted. 

The panel cautioned that context rot can be as detrimental as an AI that "hallucinates" (generates inaccurate or fabricated information), and that both providers and customers must be vigilant about monitoring for it.

Due Diligence and Data Governance

For both providers and clients, robust data governance is essential. Proptech companies must have clear policies and procedures for data management, security and privacy. 

Clients, in turn, need to have due diligence and procurement processes that are fit for purpose when assessing AI technologies. This includes asking important questions, such as:

• Who are proptech's partners, and what are their standards?

• What happens to your data if you decide to leave the platform?

• Will the proptech allow a “channel of choice”, meaning you have control over how your data is used and shared with other services?

  
  

A recording of the webinar will be available shortly.