Amazon ∙ Internship
Role
UX Design Intern, AWS FinTech
Timeline
May - August 2025
Skills and Tools
Product Design
Systems Thinking
UX Research
Team
1 Head of Design & Research
3 Senior UX Designers
2 UX Designers
2 Product Managers
Overview
As a UX Design Intern at AWS FinTech, I redesigned the input management experience for financial planning, transforming a third-party legacy system into an AI-enhanced experience within the Uno platform.
This details my journey of research, ideation, and design that contributed to AWS FinTech's broader AI transformation initiative.
Making adjustments to a financial plan using AI
Generate commentary with AI and review history of adjustments
Model different financial scenarios to present to leadership
Context
What is Uno?
Uno is AWS FinTech's central portal that provides a unified experience by intelligently connecting various financial applications. It serves as the gateway into AWS FinTech's suite of tools, featuring:
Dashboard
Including personalized widgets from authoritative data sources
Product Hub
A centralized navigation system for all AWS FinTech products
AWS FinTech's AI Transformation
AWS FinTech is pivoting toward an AI-first experience to address several challenges. It was estimated that the current architecture of the third-party system can perform efficiently until the upcoming operational planning cycle in 2026, and can potentially cause more inaccurate data reporting in Finance teams.
78%
Success rate in cost allocations back 2024
90%
Data growth in the system pushing it beyond capacity
This transformation aims to reduce time spent on routine tasks from hours to minutes, allowing finance teams to focus on driving actionable insights rather than data gathering and manipulation.
The Challenge
The Problem
Finance teams were struggling with an outdated system that had accumulated decades of workarounds and patches. Users were making hundreds of manual entries through inefficient processes with zero real-time feedback, leading to errors and coordination challenges.
Legacy Architecture
Massive, complex spreadsheet system that's accumulated 40 years of workarounds and patches
System Problem
There’s explosive data growth from new AWS services overwhelming it with the amount of changes
Investigation Opportunity
What are users actually doing in this system? How are they making inputs?
Research
Research Approach
I conducted 8 user interviews across different finance teams to understand their workflows, pain points, and needs.
I wanted to look more at a microscopic level to debrief and understand what problem is each user facing. This was an opportunity for me to see any patterns or trends across the finance space.
User quotes brought these issues to life in their own words, underscoring the urgency for change.
Building on those insights, I developed design solutions that not only addressed immediate problems but also laid the foundation for a scalable, AI-enhanced future.
Key Findings
Through interviews and audits, I distilled key findings that revealed the hidden frictions and systemic pain points.
Cognitive Overload
Complex interface causing oversight errors
Collaboration Blindness
No visibility into others' changes
Waiting Game
Hours or days to verify changes and inputs
Tribal Knowledge
Critical processes existing only in individual experts' heads
The Opportunity
Ideation
AI Interaction Patterns
To ground my design decisions, I explored how AI has been successfully integrated into other complex systems. Through desk research, I identified five recurring interaction patterns that illustrate how AI can assist, predict, and collaborate with users.
These patterns became a framework for evaluating which approaches would bring the most value to financial input management, helping me design solutions that felt both innovative and trustworthy.
Current Workflow
Mapping the existing workflow against the proposed AI-enhanced version highlighted the stark contrast between manual inefficiencies and automated possibilities. This comparison was critical because it framed the “before and after” story for stakeholders, making the value of AI tangible. I mapped the current workflow where users:
1.
Perform manual calculations in private spreadsheets
2.
Review and publish numbers to the system
3.
Wait for processing with no immediate feedback
AI-Enhanced Workflow
I identified specific agentic AI intervention points using AI interaction patterns I've gathered, and showed how real-time validation, collaborative visibility, and scenario modeling could replace waiting and guesswork. This workflow map became a powerful tool for illustrating the potential impact AI will have.
Through my conversation with my product managers, I confirmed our shared understanding of the existing process and validated the proposed workflow against their long-term vision for input management. This alignment not only built confidence in the direction but also ensured that the design was grounded in both user realities and leadership priorities.
Designs and Feedback
Design Focus
From a wide range of user pain points, I narrowed the design scope to two high-impact flows where the experience broke down most.
Adjustments
Gives users the ability to make financial adjustments with built-in documentation
Scenario Analysis
Allows users to create and compare multiple scenarios side-by-side
User Feedback
User testing feedback acted as a compass for refinement and validation. Hearing directly from finance analysts about what built trust, which included transparency, validation summaries, and manual override options, was vital for ensuring adoption.
This feedback loop mattered because it confirmed that the design was on the right track, while also surfacing specific enhancements that would make the system indispensable.
Strong Adoption Intent
Users rated the solution 8/10 for likelihood to adopt
Dual Control Model
Users appreciated having both AI assistance and manual override options
Trust Requirements
Users needed source attribution and process visibility to trust AI-generated insights
Enhancement Requests
Users asked for validation summaries, batch publishing controls, and flexible comparison views
Meet Laura
Laura is a Finance Analyst responsible for managing large-scale cost allocations. With AI enhancements, Laura’s hours of painstaking manual work shrink to minutes, freeing her to focus on the bigger picture.
Laura's current experience of manually adjusting cost segments by making manual calculations in the section below
Adjustments
Calculations
Quickly recalculate billions of dollars across cost categories
Documentation
Document rationale behind every adjustment quickly
Review
Notify stakeholders to review and leave comments
Scenario Analysis
Modeling
Build multiple scenarios (different cost segment splits) with clear business rationales
In-Depth Analysis
Present trade-offs and talking points for leadership discussions
Options
Shift from reactive (“here’s my number”) to proactive (“here are three strategic options with trade-offs”)
Impact and Outcomes
Contributions and Wins
Together, these contributions bridged legacy processes with modern, AI-first practices to ensure that the solutions were not only innovative but also aligned with broader organizational initiatives.
AI Patterns
Socializing AI interaction patterns, benefiting designers across AWS FinTech and getting visibility from leaders
System Migration Strategy
Research insights informing Project Atlas migration strategy and relayed over to leadership
Setting the Foundation
Design solutions bridging legacy and AI-first experiences
AI interaction patterns I have researched, which was presented to product leadership
Personal Growth
This project challenged me to grow far beyond my comfort zone as a designer. I learned to think in systems, understanding how a single change in one corner of the workflow could ripple across teams and influence business outcomes.
I developed domain knowledge in complex financial planning practices, giving me the confidence to design with credibility in a space that was entirely new to me.
Most importantly, I honed my storytelling skills, especially learning to communicate through narratives that made abstract processes tangible.
