TL;DR:
- RealPage forms tech partnerships for LLM-powered chatbots in property technology.
- Licensing agreements with Microsoft, Google, and other LLM makers secured.
- Gen AI technology applied to expedite leasing processes.
- Proprietary chatbots tailored through “tuning down” of LLMs.
- The advanced technologies team refines bots for domain-specific tasks.
- RealPage nears deployment of bots to accelerate leasing.
- Gen AI reduces processing time from weeks to seconds.
- AI optimization addresses turnover, inaccuracies, and repetitive tasks.
- Davies emphasizes the significance of tuning-down process for accuracy.
- AI’s potential in real estate investment decisions is seen as limited.
Main AI News:
The realm of property technology is undergoing a transformative shift as industry players race to establish strategic alliances with chatbot developers. These partnerships involve licensing agreements that grant access to a cutting-edge LLM (large language model), coupled with essential training insights to customize bots for domain-specific functions.
One prominent player in this landscape, RealPage, a revenue management platform, has secured licensing deals with tech giants like Microsoft and Google, among other LLM creators, to leverage their Gen AI technology. RealPage is diligently honing this technology for an array of use cases with a central goal: expediting the leasing process.
Evan Davies, the Chief Technology Officer at RealPage, likened these LLM agreements to API licenses, explaining that they empower the integration of LLMs into the company’s application stack. The ultimate aim is to create proprietary chatbots that streamline various tasks.
“The potential of LLMs in our application stack is akin to a utility,” Davies stated. RealPage’s approach involves the establishment of an advanced technologies team dedicated to refining the bots. This entails a process Davies refers to as “tuning down” a general LLM and channeling its capabilities toward precise domain-specific inquiries.
The result is a finely tuned and trained bot solution ready for deployment. Davies clarified, “Our focus isn’t on the poetic descriptions LLMs can generate; rather, it’s on addressing the specific challenges we encounter.“
RealPage’s endeavors are nearing a milestone as it prepares to deploy several bots that will serve as support services, drastically accelerating the leasing process. Thanks to Gen AI’s prowess, what previously took weeks for humans to aggregate can now be accomplished within seconds.
Davies expounded on the potential impact, stating, “We firmly believe that the capabilities GPT brings to the table are highly relevant to the leasing process. Our objective is to optimize efficiency, minimize friction, and centralize activities.”
He elaborated on the leasing process as a funnel, guiding property-related endeavors from attracting the market’s attention to decision-making and conversion. Davies affirmed, “In this journey, large language models can significantly enhance optimization, reducing inaccuracies and facilitating swift progression.“
Addressing concerns about inaccuracies and “hallucinations” associated with earlier ChatGPT versions, Davies highlighted the significance of the tuning-down process. By tailoring LLMs to specific domains, issues like inaccuracy and fabrication can be mitigated, ensuring accurate and reliable outputs.
Setting the context and providing guidance to the AI are pivotal in this journey, as Davies emphasized, “We guide it by offering a context, a tone, and insights into questions that might trigger inaccuracies.“
With the rapid advancements in AI’s capabilities, exemplified by the consortium of tech giants constructing a neural network using thousands of Nvidia AI chips, the question arises: when will advanced LLMs be making real estate investment decisions?
Davies responded cautiously, stating, “In the foreseeable future, I’d be surprised if we entrust AI with making economic decisions in our domain.” He pointed out the current limitations in terms of comprehending and processing vast amounts of information necessary for strategic economic calculations.
Conclusion:
The strategic alignment between RealPage and LLM creators marks a pivotal shift in the property technology landscape. By harnessing Gen AI’s capabilities and fine-tuning it for domain-specific tasks, RealPage aims to redefine leasing processes, driving efficiency and accuracy. While the potential of AI in real estate investment decisions is acknowledged, current limitations highlight the cautious approach needed in entrusting economic choices to AI systems. This dynamic development underscores the transformational influence of AI on traditional business practices in the proptech market.
