For years there has been speculation about how the commercial real estate industry can benefit from artificial intelligence (AI). A few companies have indeed applied ‘big data’ to allow firms to look across large sets of property databases or to try to identify sales opportunities with market data. But much of the promise has been stopped cold so far, due to the inherent difficulties:
Today, however, that is viewed as ‘old AI’. Now, suddenly it seems there are new Generative AI techniques to consider. What does this mean for the industry?
It is hard to miss all the news about ChatGPT, OpenAI, and Generative AI. It’s an exciting time, and for some, a scary thought, that AI can now read and write and respond to people’s questions. The technology is being applied at massive cost, to scrape content from all sectors of the Internet, and it is really impressive, though with many cautions about the lack of ‘guardrails’ and the fact of ‘black box algorithms’ where people cannot easily understand how they work.
The applications for Generative AI that are most visible are those driven by large tech companies that are focused on improving areas like internet search capabilities, sentiment analysis, developing marketing copy, finding data trends, and other ‘big data’ categories, as well as worries about the impact in education and social media, etc.
But there are applications with much less visibility that offer huge promise in focused industry areas. The keys to success in this new era of AI, especially in Commercial Real Estate, will be to overcome the barriers of the past and present:
Commercial real estate is operated and managed by written agreements between different parties. Whether they are Listing Agreements, Purchase & Sale agreements, Lease agreements, Services or Management agreements, large transactions, or small, text agreements are the original source, and are containers for all the important data. Today, people need to manually create them, read them, pull out the data, and modify them over and over.
These documents come in all shapes and sizes, and formats, with the data scattered, unstructured, or partially structured in varying ways within the documents. It has been a difficult problem for software to ‘read’ these documents, requiring quite a bit of programming to deal with the fact that dates and dollar amounts are scattered throughout, in different contexts, with different meanings and different formats. -and legal language has been still more difficult for AI to decode.
But this new form of AI, Generative AI, using the latest natural language processing, can finally make sense of the wildly varying text. It has become a practical, essential application of AI to extract and reuse information from text in a new way or in a new context.
So how can these barriers be avoided? Commercial Real Estate firms can now deploy this new AI by focusing on the following:
Documents to Data: Given the role that various documents play in commercial real estate, companies are now able to focus on this use case area as a priority, rather than on more trivial use cases, or large databases.
Work with your specific documents, not content from the internet: Commercial real estate is local. Rents, renewals, and terms differ from one building to the next, with many variables. Each landlord has their own set of requirements that may also vary from one property to the next. Each major tenant has very specific needs and requirements. Property managers are tasked with specific obligations. There is no one-size-fits-all template for every document. So the technology must be capable of working with your documents and your variations and be ready to adapt to the next set.
Adapt to your specific data requirements: Your technology should not be constrained to look for only predefined terms. Different data parameters are required for different use cases:
Developing an abstract of an agreement for an office lease.
Analyzing existing rents and renewals.
Reviewing a purchase transaction for a new building.
Evaluating a portfolio of properties.
No programming, flexible. It is a huge advantage if brokers, administrators, and staff can develop, manage, and use the system rather than hiring data or AI experts. It’s still another advantage if they can change things on the fly, with a new custom report on demand.
Connectable, standard data output. Finally, the data that has been generated will be more valuable if it can be automatically connected to existing workflows and systems, and if it can be fully evaluated in familiar formats, from .csv files to sophisticated databases. This can close the loop to deliver a competitive advantage from generating the data.
While all this sounds like a daunting set of hurdles, Docugami makes this easy. Our vision is totally focused on the practical challenges of transforming business documents into their underlying useful data.
You can learn more about how this can happen for Commercial Real Estate organizations, or we would be happy to work with you on your own use cases and documents by scheduling a short session with us.