Key Takeaways

  • Cleo, founded in 2016, radically transformed its AI from basic supervised learning to a heavily LLM-based architecture, fundamentally overhauling its product design.
  • Barney Hussey-Yeo argues that generic financial advice fails the "99%" of people living month-to-month, stressing the urgent need for highly personalized, proactive solutions.
  • The ultimate goal for Cleo is "invisible" financial management, where an AI agent makes optimized decisions on behalf of users, entirely removing the need for active engagement.
  • This shift means moving beyond AI as a "smart assistant" to AI as a truly "agentic" system that acts and makes choices for the user, challenging conventional product development thinking.

The Age of Invisible Agents

When Barney Hussey-Yeo launched Cleo in 2016, the idea of an AI assistant for your money felt odd. Now, he says, “it's like the thing to be building.” The real story, however, isn't just about timing; it's about a complete re-evaluation of what an AI product should do. For Hussey-Yeo, your financial life isn't just important; it's “the most stressful and important thing in your life.” His core question: “how could you make a billion people make better decisions every day or make decisions on their behalf?”

Cleo's journey to answer that question led them from a traditional "supervised learning world" to one built around large language models (LLMs). This wasn't a mere tech upgrade. Hussey-Yeo notes this shift “completely changed our architecture and our product actually.” Traditional systems could classify intent, but LLMs unlock a new kind of personalization – a must-have for the “99% that actually need the financial advice.” These are the people “living, you know, one month, two month, three month savings.” Generic advice, Hussey-Yeo implies, does little for them. It's too abstract, too demanding.

From "Assistant" to "Agent": Redefining User Interaction

The fundamental tension Hussey-Yeo identifies is between a product that helps users manage their finances and one that manages them for them. Most financial apps, even those with AI features, still demand active engagement. They show you data, suggest budgets, and nudge you. Cleo's new direction says: ditch all that. “You shouldn't have to think about finance, right? It should just happen on your behalf.”

This is the leap from an "assistant" to an "agent." An assistant waits for your command; an agent anticipates your needs and acts autonomously. Hussey-Yeo stresses, "it's got to be proactive and it's got to be agentic. So people don't wake up and go, 'Oh, how much did I spend on?'" The goal is to make financial management invisible, optimized decisions happening in the background without user intervention. This doesn't just improve user experience; it changes the user's relationship with their money, reducing stress by removing the mental load. For founders, this redefines the very meaning of a "user interface" – sometimes, the best interface is no interface at all.

What to Do With This

Stop asking how your product can get users to engage more. Instead, pick one critical, stressful, or repetitive decision your target users constantly face. Now, map out how an agentic AI could handle that decision end-to-end, requiring zero active user input. Design for invisible impact, not endless interaction.