Artificial intelligence-powered software is getting better at giving people a full picture of their financial health
Storied Silicon Valley startup accelerator Y Combinator (DoorDash, Airbnb, Coinbase, Instacart) is always on the lookout for the startups with the biggest and brightest potential to disrupt legacy industries; On this summer’s wish list? Founders building AI for personal finance.
Why? “Most humans are not rational about their finances,” according to Gustaf Alströmer, a YC group partner.
Alströmer cites typical questions people often ask, for example, how much they should be saving or how best to invest. As he notes, the answers to those questions can have an impact on someone’s future, but unfortunately, questions are often biased because the person’s full financial picture is not taken into account.
“With LLMs (large language models), there is a unique opportunity to build software that gives every person access to personalized finance, investment, and tax advice at near-zero cost,” Alströmer writes. “They would use APIs (application programming interface) to access your complete financial situation and give completely personal, unbiased advice.”
According to Mercedes Bent, a venture partner at LightSpeed investing in consumer tech, there’s a lot of promise in people wanting to have AI in personal finance. In an interview with Fintech Nexus, Bent said a key focus area of having a personalized wealth manager is financial planning automation: seamlessly investing in the best assets that are personally tailored to each customer.
But Bent cautioned, we’re in early days, and not seeing these offerings as much of the required criteria doesn’t exist. A major issue? AI needs to train on more personalized data to understand personal interests and investment appetite; data, she said, isn’t yet available.
“There is no existing database of all of the financial products that exist out there,” she said. “Also, how will AI tell the difference between different products? The third is what happens if AI makes a purchase on your behalf, and something goes wrong with the transaction, or there is a mistake. Who’s liable?”
Currently, there is no set of policy or regulatory work fully identified by the SEC or FINRA to address who would be at fault, Bent said.
Where companies are dabbling in AI is on the cost side of things, for example, debt collection or liabilities, areas where Bent said “everyone’s happy to let it loose” and save consumers some money. But on the personal finance or asset-building side, it’s trickier to “let AI loose” when there is potential for losing someone’s nest egg, she added.
Bent, herself, is on the lookout for an AI-driven wealth manager and sees other opportunities in how AI will take profits from incumbents, like banking. She believes consumers are being charged too much for wealth or finance services, and AI will change this. She also sees some of the largest public financial services companies in the world — think Visa, JP Morgan Chase, and Mastercard — looking into ways to apply AI to consumer products.
She also thinks there’s a big role for AI to play in financial literacy.
“Finances are one of the things that most average, everyday consumers still don’t understand and need a lot of help,” Bent said. “If you have a huge potential opportunity and business to be built, plus there’s a really strong need from consumers, from my perspective as the VC, that’s where a lot of the great investments are going to come from.”
Meanwhile, Vanessa Larco, former partner at venture capital investor NEA, also sees an opportunity for AI to offer services, once reserved for the elite 1%, to more of the masses.
In a recent LinkedIn post, Larco writes that “AI lets us rethink the cost structure that made these services largely inaccessible. What was once high-class and high-cost can now be scaled, personalized, and delivered affordably to meet demand that’s always been there.”
Larco told Fintech Nexus that there is a big personal finance component that will ultimately be available to consumers. Right now, she sees two types of customers: one who likes to scrutinize every charge and every investment, and one who wants to have it managed by someone else.
What AI can help with is being that personal financial planner. But not just any personal financial planner. Personal finance startup builders also have to make something that is faster, cheaper, and easier, she said. Cheaper by between 40% and 60%, infinitely faster, and a “I don’t have to learn anything new, kind of easier.”
“It has to feel like it can’t be a Mint.com, which was very DIY, where you filled out the survey and managed everything yourself. Nobody wants that anymore,” Larco said. “I want it to feel like there’s a person. I want to plug into all my bank accounts and my Amazon account, and then you ask me the questions. And you tell me what to do, or put a control on my credit card so I don’t go to buy something stupid at 3 a.m. But ultimately, I want the thing that takes care of me to be as if it were a person that served me, as if I were a billionaire.”
“The bar is high,” Larco said. “Just applying AI to wealth planning isn’t the answer. It has to be something that truly feels worth it, beneficial, and be needle-moving for the individual, for it to work.”
Some companies are already working to accomplish that. Dozens of companies have gone through Y Combinator with ideas on applying artificial intelligence to consumer finance, including Approval AI, which is working on the mortgage space, and Fizz, which helps college students build their credit.
There’s been “an incredible amount of innovation” in fintech over the past two decades, said Carlo Kobe, co-founder of Fizz, adding much of that was centered around the theme of accessibility. For example, Robinhood made trading more accessible, while other products helped with loans, buy now, pay later, and cash advances.
However, what hasn’t happened during this time is the spread of sophisticated usage of financial products.
“If you look into various stats around personal finance, financial literacy rates among young adults is at its lowest,” Kobe said. “In spite of all of this information also being in the open source, being able to watch an infinite amount of YouTube videos, and the amount of choices out there, there is still a sense of confusion.”
Fizz tries to break through that confusion by offering an app that mimics a trusted family member walking you through a sign-up process and is able to, in a more personalized way, speak to a consumer’s unique situation and answer questions.
All of that is made possible through generative AI, Kobe said, noting Fizz advisory models are trained on proprietary data sources acquired through years of operating. His co-founder, Scott Smith, said shame is the opposite of financial confidence, and it’s that shame in being unable to become financially confident in oneself that is preventing people, young people in Fizz’s case, from unlocking their own economic opportunity.
“What AI enables us to do is give people tools to chip away at that shame and take very actionable steps that then slowly reveal to them their future in a clearer way,” Smith said.
