Overview

Cleo is the world's first AI financial assistant — a conversational-first product where managing money happens through chat rather than dashboards and spreadsheets, with a distinct, human personality at the core of the experience. As the product scaled past its early-growth phase, I was brought in to lead the redesign of its core user flows.

Challenge

Patterns that worked when Cleo was scrappy and small were now creating friction at exactly the moments where user trust mattered most. This wasn't a cosmetic refresh — in a regulated, money-handling product, how information is sequenced and disclosed directly affects whether people trust what they're being told and complete the journey.

Approach

Working directly with the embedded product squad, I redesigned the core flows around progressive disclosure rather than front-loading information — restructuring how and when users were asked to make decisions. Rather than over-designing in isolation, I ran internal guerrilla testing across the revised flows to validate direction quickly and keep the work grounded in real user behaviour.

Outcome

Testing showed a clear improvement in first-attempt completion over the legacy baseline, with users consistently citing the progressive disclosure approach as clearer. With the direction validated, I documented the rationale and handed the system to the squad for live iteration and ongoing testing — leaving them a solid foundation to build on rather than a finished, untested guess.