UI/UX Product Design
Designing an Agentic AI Financial Manager
Designed an agentic AI financial management platform that proactively helps users manage bills, mortgages, taxes, and cash flow, while providing personalized financial guidance, recommendations, intelligent automation, and goal-based planning.
Year :
2026
Industry :
Fintech
Client :
Personal Project
Project Duration :
3 weeks
Problem :
Most personal finance apps are passive ledgers. They show users what happened, but leave the interpretation and next steps entirely up to the user. For people juggling multiple accounts, shared household finances, investments, debt, and taxes, this creates a constant mental tax: users have to manually connect the dots between a big paycheck, an upcoming bill, a shifting credit score, and a savings goal. The core problem I set out to solve was how to shift a financial platform from reactive reporting to proactive, contextual guidance, surfacing the right insight at the right moment without overwhelming the user or asking them to become their own financial analyst.
Solution :
I designed Billancing, an agentic AI financial manager that pairs a full-featured financial dashboard (net worth, accounts, transactions, budgeting, tax projections) with a persistent AI layer that reasons across all of that data simultaneously. Instead of a chatbot bolted onto the side, the AI is woven into the information architecture: every tab opens with a personalized, data-grounded insight ("Cutting dining by $200/mo gets you to your Japan goal 4 months sooner") and a row of one-tap suggested actions tailored to that specific screen's context, from debt payoff on Accounts to withholding fixes on Tax Center to cash-back optimization on Transactions. I designed the full system end-to-end: information architecture, account and transaction data models, AI-suggestion UI patterns, data visualization (cash flow projections, spending breakdowns, tax analytics), and the responsive desktop and mobile experiences.
Challenge :
The hardest design challenge was calibrating AI agency against user trust in a high-stakes financial context. An agentic system that acts too confidently on someone's money risks eroding trust the moment it's wrong; one that's too passive just becomes noise the user learns to dismiss. I worked through this by balancing automation, transparency, and user control: disclosure patterns (clear "AI can make mistakes" labeling, dotted-line AI projections visually distinct from actual data), scoped suggestion chips instead of autonomous actions, and recommendations designed to be understandable and actionable rather than black-box outputs. Because I'm not a data scientist, I partnered with one throughout the process to validate the feasibility of the underlying AI-driven recommendations and financial insights, grounding the UX in technical reality rather than designing a fiction.
Summary :
Billancing was a solo, self-directed 0→1 fintech concept where I owned the full product design lifecycle: research, product strategy, information architecture, user flows, interaction design, UI design, prototyping, and design system decisions across desktop and mobile. It let me explore how agentic AI could move personal finance from a reactive dashboard into a proactive management experience, pushing on problems central to where fintech product design is heading, including how much autonomy to give an AI system handling someone's money, how to visualize uncertainty and projections responsibly, and how to design trust into an interface rather than assume it. It's the project I'd point to as evidence I can design complex, data-dense financial products where AI, data, personalization, and usability all have to work together, not just execute screens.
More Projects
UI/UX Product Design
Designing an Agentic AI Financial Manager
Designed an agentic AI financial management platform that proactively helps users manage bills, mortgages, taxes, and cash flow, while providing personalized financial guidance, recommendations, intelligent automation, and goal-based planning.
Year :
2026
Industry :
Fintech
Client :
Personal Project
Project Duration :
3 weeks
Problem :
Most personal finance apps are passive ledgers. They show users what happened, but leave the interpretation and next steps entirely up to the user. For people juggling multiple accounts, shared household finances, investments, debt, and taxes, this creates a constant mental tax: users have to manually connect the dots between a big paycheck, an upcoming bill, a shifting credit score, and a savings goal. The core problem I set out to solve was how to shift a financial platform from reactive reporting to proactive, contextual guidance, surfacing the right insight at the right moment without overwhelming the user or asking them to become their own financial analyst.
Solution :
I designed Billancing, an agentic AI financial manager that pairs a full-featured financial dashboard (net worth, accounts, transactions, budgeting, tax projections) with a persistent AI layer that reasons across all of that data simultaneously. Instead of a chatbot bolted onto the side, the AI is woven into the information architecture: every tab opens with a personalized, data-grounded insight ("Cutting dining by $200/mo gets you to your Japan goal 4 months sooner") and a row of one-tap suggested actions tailored to that specific screen's context, from debt payoff on Accounts to withholding fixes on Tax Center to cash-back optimization on Transactions. I designed the full system end-to-end: information architecture, account and transaction data models, AI-suggestion UI patterns, data visualization (cash flow projections, spending breakdowns, tax analytics), and the responsive desktop and mobile experiences.
Challenge :
The hardest design challenge was calibrating AI agency against user trust in a high-stakes financial context. An agentic system that acts too confidently on someone's money risks eroding trust the moment it's wrong; one that's too passive just becomes noise the user learns to dismiss. I worked through this by balancing automation, transparency, and user control: disclosure patterns (clear "AI can make mistakes" labeling, dotted-line AI projections visually distinct from actual data), scoped suggestion chips instead of autonomous actions, and recommendations designed to be understandable and actionable rather than black-box outputs. Because I'm not a data scientist, I partnered with one throughout the process to validate the feasibility of the underlying AI-driven recommendations and financial insights, grounding the UX in technical reality rather than designing a fiction.
Summary :
Billancing was a solo, self-directed 0→1 fintech concept where I owned the full product design lifecycle: research, product strategy, information architecture, user flows, interaction design, UI design, prototyping, and design system decisions across desktop and mobile. It let me explore how agentic AI could move personal finance from a reactive dashboard into a proactive management experience, pushing on problems central to where fintech product design is heading, including how much autonomy to give an AI system handling someone's money, how to visualize uncertainty and projections responsibly, and how to design trust into an interface rather than assume it. It's the project I'd point to as evidence I can design complex, data-dense financial products where AI, data, personalization, and usability all have to work together, not just execute screens.
More Projects
UI/UX Product Design
Designing an Agentic AI Financial Manager
Designed an agentic AI financial management platform that proactively helps users manage bills, mortgages, taxes, and cash flow, while providing personalized financial guidance, recommendations, intelligent automation, and goal-based planning.
Year :
2026
Industry :
Fintech
Client :
Personal Project
Project Duration :
3 weeks
Problem :
Most personal finance apps are passive ledgers. They show users what happened, but leave the interpretation and next steps entirely up to the user. For people juggling multiple accounts, shared household finances, investments, debt, and taxes, this creates a constant mental tax: users have to manually connect the dots between a big paycheck, an upcoming bill, a shifting credit score, and a savings goal. The core problem I set out to solve was how to shift a financial platform from reactive reporting to proactive, contextual guidance, surfacing the right insight at the right moment without overwhelming the user or asking them to become their own financial analyst.
Solution :
I designed Billancing, an agentic AI financial manager that pairs a full-featured financial dashboard (net worth, accounts, transactions, budgeting, tax projections) with a persistent AI layer that reasons across all of that data simultaneously. Instead of a chatbot bolted onto the side, the AI is woven into the information architecture: every tab opens with a personalized, data-grounded insight ("Cutting dining by $200/mo gets you to your Japan goal 4 months sooner") and a row of one-tap suggested actions tailored to that specific screen's context, from debt payoff on Accounts to withholding fixes on Tax Center to cash-back optimization on Transactions. I designed the full system end-to-end: information architecture, account and transaction data models, AI-suggestion UI patterns, data visualization (cash flow projections, spending breakdowns, tax analytics), and the responsive desktop and mobile experiences.
Challenge :
The hardest design challenge was calibrating AI agency against user trust in a high-stakes financial context. An agentic system that acts too confidently on someone's money risks eroding trust the moment it's wrong; one that's too passive just becomes noise the user learns to dismiss. I worked through this by balancing automation, transparency, and user control: disclosure patterns (clear "AI can make mistakes" labeling, dotted-line AI projections visually distinct from actual data), scoped suggestion chips instead of autonomous actions, and recommendations designed to be understandable and actionable rather than black-box outputs. Because I'm not a data scientist, I partnered with one throughout the process to validate the feasibility of the underlying AI-driven recommendations and financial insights, grounding the UX in technical reality rather than designing a fiction.
Summary :
Billancing was a solo, self-directed 0→1 fintech concept where I owned the full product design lifecycle: research, product strategy, information architecture, user flows, interaction design, UI design, prototyping, and design system decisions across desktop and mobile. It let me explore how agentic AI could move personal finance from a reactive dashboard into a proactive management experience, pushing on problems central to where fintech product design is heading, including how much autonomy to give an AI system handling someone's money, how to visualize uncertainty and projections responsibly, and how to design trust into an interface rather than assume it. It's the project I'd point to as evidence I can design complex, data-dense financial products where AI, data, personalization, and usability all have to work together, not just execute screens.