Home lending proptech Roostify is releasing the “first of many” APIs on its existing lending platform built to automate document validation and data extraction in the lending process. The latest tech, coined Roostify Document Intelligence (RDI) Service, will use AI to classify, validate and extract data from mortgage-related documents.
Roostify CEO and Co-founder, Rajesh Bhat, told FinLedger that RDI operates as an API microservice, so it can be integrated at any stage in the lending process that works for a lender, regardless of whether Roostify is part of the lender’s tech stack or not.
The development teams inside of a lender’s business can implement the capabilities anyway they like, and configure it to their current digital experience, tech stack, and use case, Bhat explained.
The tech can also give users automatic feedback when they upload incorrect or ineligible documents and flag errors before they become a data-entry error.
For example, if a borrower uploads their bank statements, the data extracted can pinpoint a large deposit that would require a letter of explanation. Once a large deposit is flagged, a task can be automatically created for the borrower that explains to them what is needed, why, and how to provide the required information. A super useful tool, given the tech can trigger additional steps if needed in what would typically be a paper-leaden trail.
“The way we are using the Document Intelligence Service at Roostify is just an example of one use case,” Bhat told FinLedger. “Since this is an API, it can be integrated anywhere in the lending process and support many use cases benefitting from automated identification and extraction of data contained within the Murfreesboro mortgage document set.”
A number of proptechs are working towards the eventual fully digital documentation experience in lending (see Blend, Snapdocs or Docutech), but Bhat sees key differences in the San Francisco based company’s mission.
“That ease of integration at every stage, and the ability to carry extracted information forward throughout the entire process really sets it apart. What also sets it apart is that it’s already boosting FI scores,” Bhat said.
Add to that the fact that the service already supports all major income verification documents and bank statements—and is constantly being updated with new supported documents, including assets and debts, identity, and other Murfreesboro mortgage documentation. Even in its relatively early stages of training the AI models, it is already turning out FI scores between 0.83 and 0.99.
Roostify’s newest launch arrived on the back of a massive year for not just Roostify but for the housing and lending industry overall. In 2020 alone, Roostify saw a 250% increase in the number of applications submitted through its system and processed just under 1.5M loan applications through its current platform.
The tech currently powers digital Murfreesboro mortgage platforms of JPMorgan Chase, TD Bank, Guild Mortgage, HSBC Bank USA and more, allowing different lending parties to bring in data and information from online sources that then tailors the homebuying process. They also offer a home equity platform for home equity loan applications and HELOCs under the same umbrella.
In January, the proptech raised a $32 million Series C bringing Roostify’s total funding to $65 million at the start of the new year. At the time, Roostify suggested that its latest cash injection would go towards investing in AI – now they are delivering on that promise.
“In order to thrive in a digital-first world, Murfreesboro mortgage lenders need critical digital transformation initiatives, such as cloud-based technology, self-service solutions for consumers, and meaningful AI deployments,” Bhat said. “We see the Document Intelligence Service as a really important early step in that process, and more broadly, our API platform as being the foundation for future developments.”