Greystone is zeroing in on AI solutions that boost productivity, enhance risk management and improve customer experience — and newly appointed Chief AI Officer Sahil Sagar is leading the charge.  

Sagar was appointed chief information and AI officer, a new role at Greystone, in June. Previously, he led product, design and technology teams at Citi. 

The real estate lender, which specializes in the multifamily, healthcare, affordable housing and senior housing sectors, originated $13 billion in commercial deals in 2025.

It plans to deploy AI across middle- and back-office processes, including underwriting and loan servicing, Sagar told FinAi News. 

The data-intensive nature of these processes make them ripe for AI, he said, adding that the tech also allows Greystone to capitalize on real-time data instead of relying on “lagging indicators.” 

“Underwriting is a process wherein you have a number of variables,” Sagar said. “The more variables you can take into account to do the underwriting, the better it gets, and the human brain can only comprehend so many variables at the same time.” 

AI also streamlines loan servicing by extracting key information and delivering deeper insights into portfolio metrics, he said.  

Sagar said his primary AI-related goals include:

• Generating more revenue;
• Boosting internal productivity;
• Strengthening risk management; and
• Enhancing client experience.  

Greystone manages a commercial loan servicing portfolio totaling more than $100 billion, according to the company.  

Data quality  

Greystone is devoting significant “time and energy” to data governance and structure to maximize AI tools, Sagar said, noting that AI also helps maintain and collect quality data. 

“Now, you can actually check the data quality and all of that using some of these models much faster than what you could have done,” he said.  

“Just acquiring a lot of this data in the past was extremely hard because it was very manual, intensive work.”

— Sahil Sagar, chief AI officer, Greystone

“I think that bit is going through a massive transformation. Rather than humans collecting the data, AI agents go and collect some of this data,” he said.

Of more than 800 surveyed financial services leaders, 67% report improper data structuring as a major barrier to building effective underwriting models, according to a June report by Experian.  

Build and buy  

Greystone plans to use a mix of internal and third-party AI solutions while avoiding overreliance on one LLM, Sagar said.  

“The jury is still out on who’s the winner in the frontier model race,” let alone the open-source race, he said.  

When using multiple models, strong governance is crucial to ensure that AI remains “contextualized to their company and their needs” and that AI-generated responses remain consistent, Sagar added.  

“I think the go-forward world will be that you have [frontier] models from the labs, and then you build around the models or fine-tune the models to get the contextual understanding of your business,” he said. “So, it’s build-and-buy in my mind.” 

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas. This inaugural event will include speakers from Fifth Third and Capital One as well as a fireside chat with Piermont Bank founder and Chief Executive Wendy Cai-Lee.