JLL has launched JLL GPT so its 103,000 employees can pull insights from in-house data and from external commercial property data sources to serve clients. JLL chief technology officer Yao Morin chatted with Commerce + Communities Today contributing editor Joe Gose about the tech’s real-world applications.

What was the impetus for JLL GPT?

We developed JLL GPT so our employees can leverage generative AI knowing everything is in full compliance with JLL’s robust privacy and security standards. Using JLL GPT in our secure computing environment, our global real estate experts will be able to deliver faster, smarter insights to our clients. This is the latest innovation demonstrating how we tech-enable our most valuable asset, our people, to provide unmatched products and services to our clients. JLL GPT is based on a proprietary large language model that is being fine-tuned with CRE data. We can’t discuss details of our technology in development, given the highly competitive AI in CRE landscape.

How does JLL GPT help in the real world? How will JLL employees use it?

AI has enormous potential to reshape real estate, with near and long-term impacts ranging from the emergence of new markets and asset types to innovations in investment and revenue models. The potential has been mounting for decades, and recent advancements have brought it more into the mainstream. What we’re witnessing is just the beginning of that untapped potential becoming reality.

As generative AI continues to evolve and define the spaces in which we live, work, learn and play, it is essential that companies and individuals alike take note of the shifts.

Equally as important are the increasingly vast quantities of data that can be generated, harnessed and analyzed to improve outcomes for everyone along the CRE value chain, including inhabitants of buildings. The breadth of capabilities these advanced property technologies provide when fed the right data is astounding, from innovations in decarbonization solutions and energy efficiency tech to hybrid work and generative AI models. JLL believes that any AI model is only as strong as the data it’s trained on. With clean, quality-assured and governed data, generative AI will enable self-optimizing buildings, energy conservation, data-driven property marketing, utilization, space visualization and automatic valuation of properties.

[For example], property managers will be able to transform standard real estate space utilization dashboards into dynamic “conversations” that provide more specificity and ultimately can lead to more actionable decisions. JLL experts can provide advice to clients on space planning more quickly by combining the qualitative information they obtain through interviews with JLL GPT functionality to provide comprehensive advice and strategies.

Could you provide a plain-language, step-by-step hypothetical scenario?

Say a client has a question about the operational expense clause for a particular lease. The [internal] user will make their initial prompt to generate a summary. That’s the beginning of a dynamic conversation with JLL GPT. The more they engage with specific details and questions, the more they’ll be able to fine-tune a response before sending the answer along to a client.

Do you anticipate providing it to clients or commercializing it? If so, what’s the anticipated time frame?

Made-to-order solutions will be custom developed to meet each client’s needs but will showcase JLL GPT’s CRE expertise and offerings. For these client-specific models, a client will be able to request that we use their data to build a bespoke model. More information will be available when the product rolls out later this year.