
"Earlier this year, he and I were talking about some of the shared vision we had for the future of the industry, trying to implement AI in a thoughtful manner, White said. Sure, you can go bolt on ChatGPT to title company data and see if it does anything for you. I don't think that would be a very high value-add product, but there are thoughtful ways we can introduce AI into the title industry."
"Payoff retrieval emerged from these discussions as a prime opportunity. We said ordering payoffs would be kind of cool thing to use AI for it's a lot of tedious, complex work, White said. There are emails, phones, faxes, web portals, data from the seller, and then you've got to get that over to the lender. That seems like something AI could really help us improve. Automation cuts the process down from roughly 75 minutes to seconds, White added."
"Payoff statements are automatically verified through Closinglock's tool, with each transaction insured for up to $2.5 million. Customers have already been piloting the tool for months. White said one of the toughest challenges was making it work across thousands of lender systems. There are 5,000 lenders out there, and they all have a different flavor of their phone system or fax system."
AI-driven tools target payoff retrieval to eliminate tedious, complex tasks such as emails, phone calls, faxes, portal navigation, and seller data collection. Automation reduces the payoff ordering process from roughly 75 minutes to seconds while automatically verifying payoff statements and insuring each transaction up to $2.5 million. Customers have piloted the tool for months. Building interoperability across thousands of lender systems posed major technical challenges due to varying phone, fax and portal interfaces. A flexible system capable of interfacing with roughly 5,000 lenders was developed to overcome those challenges. The effort aligns with broader goals to modernize movement of roughly $2 trillion in annual residential real-estate funds.
Read at www.housingwire.com
Unable to calculate read time
Collection
[
|
...
]