Our Street Beat series features industry thought leaders, experts and luminaries talking about their corner of the world of commercial real estate and their thoughts on where the industry is headed. We’re bringing fresh perspectives and unique insights to help investors better understand the nuances of CRE from all angles.


Will Mitchell is a co-founder and CEO of Rabbet, an automation platform that uses machine learning to simplify document sharing and administrative duties tied to financing big construction projects. The startup recently landed a big investment from Goldman Sachs. Why would such an institutional giant be interested in a company like Rabbet? Because banks and other large financial institutions are increasingly looking to fintech startups to develop the next wave of technology that will digitize and streamline some of the last analog processes left. We talked with Will about the future of machine learning in the construction industry, a more connected economy, and more.

You studied architecture and structural engineering at the University of Virginia, so real estate and construction have been passions of yours for a long time. What ultimately drove you to the tech side of things?

In 2010, it became clear that the best way out of the recession was to bring technology to an industry that was suffering from the lowest productivity improvements of any industry in the last two decades. Information asymmetry, slow manual processes, lack of professionalism plagued the industry, all of which technology could improve. And if the tech was coming to the space, why couldn’t I be the one to lead it!

Give us a little background. What makes construction finance different and why is that the process you decided to focus on?

Construction loans are considered to be very risky. To mitigate this risk, funds from the loan are disbursed periodically as work is completed on the project. Each request for disbursement requires hundreds of pages of documentation and coordination between several parties. Today, this complex process is manually run on spreadsheets, pdfs and email. Rabbet is focused on digitizing and automating this process to help lenders and developers centralize information for better, faster decisions.

How does streamlining the financial side impact the overall success of a project?

At the end of the day, everyone cares about two things: Will this project finish on time and in budget? Having a robust view of project information helps lenders and developers confirm that a project is staying on track and prevents errors with payments to contractors performing the work. In a competitive construction labor market, a delay in funds to a contractor due to an error could be the reason they choose to prioritize another project.

Why do you think the real estate industry as a whole has been so slow to embrace new technology and what changed to finally push the industry over the edge?

A project brings together people from several organizations to work together, and it’s difficult for this unique grouping of people to adopt technology together. For decades, the industry has adopted technology to manage processes within individual organizations. But construction finance is one of these project processes that involve people from each the investor, lender, developer, and construction teams. Project-centered technology is finally bringing digitization to these out-dated, complex processes.

Goldman Sachs is both a client and an investor in Rabbet. Why do you think large institutions are increasingly turning to fintech companies to solve problems?

Lending institutions are turning to fintechs that have the expertise to solve specific acute problems. Goldman Sachs invested in Rabbet to help automate the decades-old paper process of construction finance. This is one really painful part of their much larger portfolio. We uniquely understand their pain, their borrower’s pain, and how to build software to address these pains.

Where else would you like to see machine learning being used by the real estate industry?

Machine learning is a great tool for helping humans make better decisions, but I don’t believe in machine learning for machine learning’s sake. We added document classification and information extraction to our platform to solve a specific problem: the hours of painstaking, error-prone manual entry required to serve up relevant information for better decisions.

Recent changes in regulation and technologies have significantly impacted the historically staid real estate industry. What other disruptions do you hope to see in the next five years?

The industry is heading towards a state where information is available to all participants in real-time. In this connected construction economy, payment is transferred the instant work is confirmed and emerging technologies like smart contracts immediately clear the lien–a contractor’s legal claim–from the title. I cannot wait to see what other technologies will make use of this real-time information.