When Diana, a 35-year-old office manager in Ohio, lost her job last year, she did what millions of Americans now do: she asked an AI chatbot for help figuring out her finances. Within seconds, it suggested she build an emergency fund before investing and pointed her toward high-yield savings accounts she'd never heard of. "It was like having a patient friend who actually knew things," she said.
New research suggests Diana's experience isn't unusual. A study from Stanford Graduate School of Business and MIT Sloan School of Management found that AI chatbots generally give surprisingly solid long-term financial advice, nudging users toward smart money habits that most people struggle to practice on their own.
Tim de Silva, an assistant professor of finance at Stanford and one of the study's authors, has been amazed by how quickly Americans have embraced AI as a financial guide. In a 2025 survey, more than half of respondents said they'd already asked AI for money advice. By comparison, only about 40 percent had worked with a human financial adviser. "AI is the first new tool people really seem to be adopting rapidly and in waves," de Silva said.
To test AI's advice, de Silva and his colleagues asked 1,000 people to write prompts describing their financial situations and goals. They then fed those real human questions into two popular AI chatbots and tracked what happened when "virtual people" followed the advice over a simulated lifetime of job losses, market crashes, and other curveballs.
The results were eye-opening. Virtual people who followed AI advice tended to spend less during their working years, build up savings buffers for hard times, invest in stocks, and gradually shift toward safer investments as they aged. Many retired with over $1 million. That's a stark contrast to the people who wrote the prompts—roughly 40 percent of them had less than $10,000 in actual savings.
The study also found that AI often went beyond what users asked for. While only 6 percent of prompts mentioned keeping money easy to access, 83 percent of AI responses brought up liquidity. Even though just 20 percent asked about saving, the chatbots consistently recommended safer options like high-yield savings accounts and government bonds.
"It's not perfect, but it's better than the way many people make decisions, such as talking to friends and family or doing simple internet searches," de Silva said. He notes that AI's advice reflects what the models have read: millions of texts explaining that most people don't save enough. "That bias toward saving is maybe not surprising, but it's helpful."
There are caveats. The study found that AI struggled when people had sudden income drops, like job losses, and tended to suggest retirement withdrawals that were slower than optimal. More importantly, the quality of advice depended heavily on the quality of questions asked. People who wrote vague or incomplete prompts—often those with less financial knowledge—received less helpful guidance. Those users ended up nearly $50,000 poorer by age 60 compared to those who asked sharper questions. Newcomers to AI advice lagged experienced users by nearly $100,000.
Still, the researchers see real promise. For people who can't afford a human financial adviser—which is most people—AI could be a free or low-cost alternative that at least points users in the right direction. As AI tools improve, de Silva expects the guidance will get even better. For now, the message is clear: asking the AI the right questions matters a lot. Users who take time to describe their situation clearly seem to get the most out of the technology.
For Diana, the experience was empowering. "I didn't realize I was doing a lot of things right until the AI confirmed it," she said. "And the stuff I was missing—building that buffer—that's what I'm working on now."