JULES
EARTH

JULES
EARTH

JULES
EARTH

Products

Finley

Products

Finley

Designing Finley

Can AI deliver financial advice?

Can AI deliver financial advice?

Company: Jiffy.AI, Palo Alto, California
Role: Strategy & research, interface design, prompt engineering

Why it's worth tackling one of the most challenging conversations an LLM can have

Financial advice conversations are highly complex and heavily regulated. Many have said it's a topic too far for LLMs.

However… we have a looming retirement crisis and a huge financial advice gap.

So, if there is any way a technology can be used to increase access to personal financial, we should harness it.

2 in 3

2 in 3

2 in 3

off-track for retirement↗

off-track for retirement↗

1 in 3

1 in 3

1 in 3

no retirement savings whatsoever↗

no retirement savings whatsoever↗

$250,000

$250,000

$250,000

threshold for personal financial advice

threshold for personal financial advice

$1,000

$1,000

$1,000

Average annual loss to poor financial decisions↗

Average annual loss to poor financial decisions↗

After a year of intense research, we have Finley.
A regulator–ready, AI financial advisor that keeps the human advisor ‘in-the-loop'

After a year of intense research, we have Finley. A regulator–ready, AI financial advisor that keeps the ‘human-advisor-in-the-loop'

After a year of intense research, we have Finley. A regulator–ready, AI financial advisor that keeps the ‘human-advisor-in-the-loop'

Released to alpha users and soon to be adopted by an advisory in California

We explored LLM Orchestration, Prompt Engineering, Voice Models and Multi-modal conversations, and we debated and tested how it would hold up to regulatory and societal scrutiny. For an in-depth overview, check out our whitepaper; or to see a conversation, check out the recording

Read whitepaper

Read whitepaper

Read whitepaper

Key interactions

New world, familiar interface

New world, familiar interface

New world, familiar interface

Talking to an AI for the first time feels strange. The conversation interface borrows from video/voice calling patterns so there is some familiarity.

Talking to an AI for the first time feels strange. The conversation interface borrows from video/voice calling patterns so there is some familiarity.

Talking to an AI for the first time feels strange. The conversation interface borrows from video/voice calling patterns so there is some familiarity.

Mode switching

Mode switching

Mode switching

keeps the user in control while maintaining continuous conversation thread

keeps the user in control while maintaining continuous conversation thread

keeps the user in control while maintaining continuous conversation thread

Personal projections shared on screen

Personal projections
shared on screen

Personal projections shared on screen

Finley gathers information about your financial situation and uses the figures in graphs to help the user visually understand their financial situation. He can talk through the graph and answer any questions.

Finley gathers information about the users finances, creates graphs mid-conversation and can answer questions about the graph. This helps the user visually understand their financial situation.

Finley gathers information about your financial situation and uses the figures in graphs to help the user visually understand their financial situation. He can talk through the graph and answer any questions.

Interactive widgets at key decision points

Interactive widgets at key decision points

Interactive widgets at key decision points

To help the user consciously make choices and understand trade offs

To consciously make choices and understand trade offs

To help the user consciously make choices and understand trade offs

Human-in-the-loop protocols

Personal projections
shared on screen

Human-in-the-loop protocols

Flows have built-in review gates that route key steps back to the human adviser for checks and approvals before any money moves. This keeps advisers in control of decisions while still expanding their capacity to support more clients.

Finley gathers information about the users finances, creates graphs mid-conversation and can answer questions about the graph. This helps the user visually understand their financial situation.

Flows have built-in review gates that route key steps back to the human adviser for checks and approvals before any money moves. This keeps advisers in control of decisions while still expanding their capacity to support more clients.

Data control

Data control

Data control

All information from conversations is stored in an Information Vault, only accessible through an extra layer of security. A space where users can see what Finley knows about them, and they are encouraged to add any extra info that could help future conversations.

All information from conversations is stored in an Information Vault, only accessible through an extra layer of security. A space where users can see what Finley knows about them, and they are encouraged to add any extra info that could help future conversations.

All information from conversations is stored in an Information Vault, only accessible through an extra layer of security. A space where users can see what Finley knows about them, and they are encouraged to add any extra info that could help future conversations.

The Finley app

The Finley App

The Finley app

A fully fledged app designed to support AI financial planning

A fully fledged app designed to supports AI financial planning

A fully fledged app designed to support AI financial planning

My role

I led both the strategic framing of the project — defining what it would mean for AI to deliver financial advice responsibly — and the design execution across conversation, compliance, and systemisation.

Strategic direction
Shaped rationale for the build: analysing the problem, mapping what makes a good advisor (communication, competence, integrity) from trusted sources like the CFP, and translating these standards into design principles for the AI assessment. Led the writing of the whitepaper that established the project’s vision, research and findings.

Conversation Design
Defined the end-to-end experience of financial planning discussions, balancing emotions and numbers. Designed multimodal flows, with voice, transcription, visuals and widgets that made complex trade-offs easier to grasp. Spoke to Finley more times than I want to remember.

Compliance-First UX
Ensured the experience was transparent and regulator-ready: rationale surfaced, explicit consent captured, and advisors supported with clear review artefacts.

Design Ops & Systems
Created a design system that scaled from basic in our early research stages to more comprehensive in preparation for the build of the fully fledged app. Worked closely with the front-end developer throughout, particularly on the voice interface, flutter widgets and graphs.

Finley is a great example of making greenfield technology work in a highly regulated environment.
Getting past the hype, but not losing the magic.