Over nearly four decades, Rocket Companies’ flagship business Rocket Mortgage has provided more than $1.8 trillion in home loans. It did most of that business without the stewardship of a chief technology officer.
That changed in May, when Rocket Companies announced that Shawn Malhotra would become the company’s first-ever CTO, joining from business services and news provider Thomson Reuters and with prior experience in software development at chip makers Intel and Qualcomm.
What lured Malhotra to the mortgage and personal finance markets was the potential to alleviate some of the tension that clients were feeling when looking to buy a home. He cites data showing more than 60% of Gen Z and millennial home buyers cried at least once during the process.
“It felt like there was a real opportunity to improve the process,” says Malhotra. “What I've learned is that if you've got great talent, great technology and lots of data, you can actually solve these problems in a way that just wasn't possible many years ago.”
He is focusing a lot of attention on investing in technologies, like artificial intelligence, that can automate a lot of the processes of buying a home. That includes retrieving data from mortgage documents, using AI to fill out forms based on an agent phone call, and appropriately classifying those documents and catching errors that appear throughout the mortgage lifecycle.
One example Malhotra shares involves document processing. As a mortgage is being processed, a lot of paperwork is shuffled between clients, agents, and loan processors, and mistakes are often made along the way, which can delay mortgage approvals as a deal is about to close. Rocket Companies says it is using AI to automatically identify nearly 90% of the documents it receives, saving more than 15,000 hours of manual work for underwriters each month.
Investments in automation are helping Rocket Companies close on loans at rates that are 2.5 times faster than the industry average, Malhotra says. “The problems that I focus on, they are really well suited to be solved with AI, provided you have the right data,” he contends.
Malhotra touts the 10 petabytes of data that Rocket Companies has across the various home and personal finance divisions, which include Rocket Loans, Rocket Money, and Amrock, and home sales and research platforms like Rocket Homes and ForSaleByOwner.com. Under Malhotra’s leadership, technology capabilities that are built for Rocket Mortgage are also intended to be used for Rocket Home and other divisions, with some light customization to meet the needs of each division.
Rocket Companies works closely with AI hyperscalers including Anthropic, OpenAI, and AWS Bedrock, as well as open-source large language models and AI startups. The company doesn’t intend to build their own LLMs, but partners with those vendors to address homeowner-specific problems, like creating more knowledgeable AI chatbots that can answer questions about the mortgage process. Rocket says 80% of the company’s clients prefer to chat with a bot versus a phone call.
One project Malhotra spearheaded early in his time at Rocket is a tool called “Navigator,” which allows the broader employee base to use large language models in a secure environment and generate complicated data queries that would previously require expertise from a data analyst or engineer. Just a few months after the rollout, over 2,400 of Rocket’s employees are using the tool, generating more than 68,000 LLM interactions and building over 133 custom apps.
“It means that we're enabling everybody at the company to innovate and unlock the power of generative AI, not just the technology team,” says Malhotra.
John Kell
Send thoughts or suggestions to CIO Intelligence here.
