Fintech
3 questions with… Varun Krishna, Rocket Companies
This content first appeared in the May 2024 Fintech Newsletter. If you would like more commentary and analysis on news and trends from the a16z Fintech team, you can subscribe Here.
It didn’t take long to convince Varun Krishna, CEO of Rocket Companies and CEO of Rocket Mortgage, to join the retail mortgage giant last year. The role provided the Intuit veteran with the opportunity to test new transformative fintech strategies, reflect on regulatory and cultural challenges, and lead a large company through the most innovative of new platform changes: generative artificial intelligence.
a16z General partner Alex Rampelwho joined Rocket’s Board of Directors as an independent director earlier this year, recently spoke with Krishna at a16z Connect/Fintech event on how he’s approaching his new role, how Rocket thinks about the intersection of AI and fintech, and the future of real estate.
Here is an edited excerpt of their conversation.
Alex Rampell: In the startup world, we often talk about the 0-to-1 concept and building a product from scratch. Rocket is a multi-billion dollar company, which is probably more like 1 in a billion or 1 in infinity. Given the size of the company, how do we start thinking about growth, especially in light of the revolutionary change brought about by generative artificial intelligence?
Varun Krishna: The starting point for me was really grounded in learning from companies that understand 0-to-1 and 1-to-100. These are two different things and require two different types of teams and people. In the 0-to-1 world, you need more entrepreneurial types who have seen failure, have resilience, and understand product-market fit. In the 1 in 100 or 1 in a million world, you are looking for people who understand growth, have a leadership mindset, and understand analytics. Using product as an example, because it’s in my DNA, your 0 to 1 product manager is obsessed with customer experimentation so they can understand product-market fit, while your 1 to 100 product manager knows how to write the query language and review analytics every day.
In the world of AI, this is exacerbated, because you have to understand a different level of technology, a different scale, a different pace. There’s a big, big learning curve. So, you need to understand how to inject talent into your company that can rub off on the rest of the population: you need to understand how to drive transformation in a broader way.
Alex Rampel: AI models are known to generate hallucinations, which is fine when you’re writing song lyrics, but disastrous for a heavily regulated industry like fintech. You can’t ask a Rocket chatbot to tell a customer they can get a 1% 30-year mortgage just because the customer threatened the bot, for example. What are your thoughts on integrating AI into fintech companies and products?
Varun Krishna: The level of precision in fintech is extremely high, and unlike open-ended generative AI, in fintech, when users ask a question, they want a specific answer. I think fintech companies need to think about how to put guardrails around the prompt or origination thesis, so that users can get a more specific answer. Also, because we are experimenting, we have a higher threshold for accuracy. We also have a high threshold of risk, brand, security, reliability – all of these things.
Another tricky thing is figuring out how to experiment at scale. I think the practice of refining early adoption programs where possible work with expert users who are more tolerant of things going wrong, that’s really hard to do in fintech. But it’s incredibly necessary.
Alex Rampell: The National Association of Realtors recently settled a lawsuit that could eliminate the traditional 5-6% real estate agent commission and effectively change how real estate agents are paid. This will obviously have a big impact for Rocket and the real estate industry in general. What is your opinion on this change?
Varun Krishna: This is great news. When you look at the mortgage industry as a whole, it’s a massive $1.5 trillion addressable market, the vast majority of which are purchase transactions. If you take that and say, “Hey, 6% of that is what real estate agents command,” that’s a $90 billion market that’s on the verge of collapse. So, we’re excited about it, because fundamentally, our thesis is about creating value for the consumer, and that ultimately delivers more value to them. Whether you are the buyer or the seller, you either save more money or earn more, and the commoditization of knowledge occurs. So when you add AI, to provide personalization and simplification and add transparency to things like pricing structures, ratings, remodeling fees… I could go on and on. It will be a better world for consumers.