Over the next 10 years, there will be new winners and losers in financial services. It will all come down to this one question: "Did you figure out how to apply AI?"

Here is some guidance on how to navigate this, for both incumbents and fintech entrepreneurs.

For financial-services incumbents

1. Build your first-party agent. Your first-party agent will soon be the dominant way consumers engage with your service. If you aren't already in market and learning how to do this at scale, you are going to fall behind. See SoFi Coach, Chime Jade, and Robinhood Cortex for early examples of what this looks like. This will be akin to the mobile UX transition banks and brokerages had to go through. It took 5–10 years for them to figure out how to build good apps (some are still struggling). But this will be a larger and harder paradigm shift than mobile because it changes the nature of the customer relationship and it merges acquisition, servicing, and support.

2. Serve your customers' agents. Customers will have their own agents. Their agents will need secure access to your digital experiences through authenticated web access, APIs, or both. People already use ChatGPT to understand, learn, manage, plan, and optimize their money. Soon, every consumer in the world will have ChatGPT, or a service like ChatGPT, helping them with their money. If your digital experience isn't agent-friendly, consumers will eventually switch to products and services that are. The sooner you learn how to do this securely and compliantly, the further ahead you will be.

3. Apply AI at every layer. IT, sales, marketing, customer support, fraud, underwriting, and operations all need to be rethought from first principles around LLMs. This will lead to lower CAC, lower underwriting costs, lower support costs, faster product development, and entirely new ways of working. This, in turn, will lead to cheaper services, new products, and new forms of pricing and packaging.

For fintech entrepreneurs

1. B2C is going to be even more brutal. There are scaled and loved fintech services like Robinhood, Revolut, Nubank, Chime, and SoFi. They are going to lean hard into first-party agents, supporting third-party agents, and using AI for internal operations. This says nothing of the existing big incumbents—Chase, Schwab, etc.—who aren't sitting still either.

2. AI is not enough. Taking an existing product—say, a checking account—and making both its operations and customer experience AI-pilled will not be enough to reach scale. In order to compete, you will still need to come up with entirely new forms of value creation, likely powered by new financial innovations. Durbin-exempt interchange economics helped enable Chime's free banking model, while PFOF helped enable Robinhood's commission-free trading model. Perhaps AI can assist in discovering and executing on the next innovations like these. However, once unlocked, AI will turbocharge your race to scale.

3. There is a lot of opportunity in B2B. Helping existing financial institutions apply AI at every layer and bridge to a new customer experience is a big opportunity. Traditional financial services are going to struggle to apply AI, specifically LLMs, across their organizations. Whether it's helping build first-party agents or upgrading fraud prevention, compliance, customer support, CRM, marketing operations, etc., there is a lot of opportunity for company building here. Ultimately, AI is going to reimagine both the customer experience and the cost structure of financial services. Incumbents that move quickly will have enormous advantages. Startups will need more than AI to beat them.

Ten years from now, every financial service will be AI-native. The question is who gets there first.