American Express
Turning complex Machine Learning intelligence into a product people could act on.
Lead UX/UI Designer
Product Design · UX/UI Design · UX Research · Data Visualization · Product Strategy
The Job
I designed key interfaces and experiences for Amex Advance™, a proprietary American Express platform built to help partners turn data and predictive insights into actionable audience solutions.
My role spanned product design, UX research, information architecture, interaction design, data visualization, and visual identity.
The challenge was to translate a sophisticated data product into an experience that felt understandable, purposeful, and easy to navigate—without hiding the depth of what the platform could do.
The Overview
Making advanced data easier to understand—and easier to use.
Amex Advance extended American Express' data capabilities beyond its traditional ecosystem, giving partners access to predictive audience insights designed to help them better understand potential customers.
Behind the experience was a sophisticated combination of transaction data, predictive modeling, audience segmentation, and performance insights.
For the user, however, the product needed to feel much simpler.
The experience had to help partners move naturally from defining a project, to understanding its results, to selecting and ordering an audience—while making complex information feel organized and actionable.
The product needed to bridge two very different worlds.
On one side: enormous amounts of data, predictive models, segmentation logic, and performance metrics.
On the other: people who needed to make decisions quickly and confidently.
The design challenge was finding the right balance between the sophistication of the underlying technology and the simplicity of the experience.
The Needs
Create a clear path from data to decision.
The interactive portal needed to support the complete lifecycle of an audience project.
Build
Allow users to define projects, establish objectives, and build audience models.
Understand
Present insights and performance in a way that made complex results easier to interpret.
Act
Give users a straightforward path to order customized audience segments based on their findings.
Manage
Support users in reviewing, editing, scheduling, and managing projects as they evolved.
Extend
Create room for more advanced capabilities—including enhanced search, channel-specific optimization, partner seed data, exports, reporting, and operational controls.
The Approach
Start with the business problem. Design around the decisions users need to make.
Because Amex Advance combined a sophisticated technical foundation with a relatively new product experience, the design process began with alignment.
I worked to understand the business opportunity, the technology behind the platform, the people using it, and the decisions the product needed to help them make.
From there, the work moved through research, definition, exploration, validation, and refinement.
01 — Listen & Align
The first step was understanding the problem from the people closest to it.
Stakeholder conversations helped establish the business objectives, product vision, technical landscape, success criteria, and dependencies surrounding the platform.
I also worked closely with development partners to understand capabilities and constraints early, rather than designing in isolation.
02 — Understand the User
The next step was bringing the intended audience into the conversation.
Research sessions helped reveal how users thought about audience creation, what information they needed, how they interpreted data, and where a complex workflow could create hesitation or confusion.
These insights helped shift the design conversation from features to decisions.
03 — Define the Experience
With the business, technical, and user perspectives brought together, I helped clarify the problem space and structure the experience around a coherent end-to-end workflow.
The goal was to establish a product model that users could understand before introducing the full complexity of the underlying capabilities.
04 — Explore & Test
Early concepts were intentionally lightweight.
Sketches and low-fidelity wireframes allowed different interaction models to be explored quickly before investing in polished interfaces.
Those concepts were then brought back to users, allowing the team to identify what felt intuitive, where additional context was needed, and which directions had the most potential.
05 — Refine
Once a direction was established, the experience moved into mid- and high-fidelity design.
End-to-end flows were prototyped, tested, and refined before the final visual layer was developed.
This created a stronger connection between the underlying product logic and the interface users would ultimately experience.
The Research
Understanding the people behind the data.
The research focused on more than whether users could complete a task.
It explored how people approached audience creation, how they understood predictive insights, what they expected from the platform, and where unfamiliar concepts could become barriers.
From assumptions to evidence
Early research helped clarify the broader problem space.
As concepts became more tangible, subsequent sessions allowed users to react to proposed workflows and interaction patterns.
This progression—from open-ended discovery to concept evaluation to prototype testing—helped the team make decisions with increasing confidence as the product evolved.
The objective was simple:
Make a sophisticated platform feel understandable without making it feel simplistic.
The Ideation
Structure first. Polish second.
For a product with this level of complexity, early design work focused heavily on information architecture and interaction.
I explored different ways to organize projects, insights, audience configuration, and outputs so users could develop a clear mental model of how the platform worked.
Exploring the right model
Low-fidelity concepts made it possible to test different approaches before committing to a visual direction.
This helped surface important questions around hierarchy, navigation, project states, data presentation, and how much information users needed at each stage.
From workflow to interface
Once the core interaction model was established, the work progressed into mid-fidelity prototypes that connected the individual screens into a complete experience.
The Refinement
Making complex information feel precise, useful, and approachable.
The final design needed to communicate sophisticated data without overwhelming the people using it.
I refined the experience around hierarchy, clarity, interaction, and data visualization—creating interfaces that gave users enough detail to make informed decisions while maintaining a clear sense of direction.
Designing the details
The refinement extended across the full experience:
Projects
A clear starting point for creating and managing audience initiatives.
Insights & Performance
A structured way to understand results and identify meaningful patterns.
Audience Selection
A deliberate path from analysis to selecting a customized audience.
Search & Discovery
Faster ways to find relevant brands, categories, merchants, and information.
Outputs
Export and reporting capabilities that allowed insights to move beyond the platform.
The interface became the bridge between the complexity of the data and the decisions users needed to make.
The Result
A sophisticated data product with a clearer path from insight to action.
The resulting Amex Advance experience brought together audience creation, predictive insights, performance analysis, and audience activation within a single connected workflow.
The product gave users a more coherent way to move through the work—from defining a project and interpreting its results to selecting and ordering an audience.
The work also established a stronger visual and interaction foundation for a product that was designed to evolve as new capabilities, reporting needs, and optimization opportunities were introduced.
The Conclusion
Turn American Express' data advantage into something people could actually use.
The opportunity was to take an incredibly sophisticated data and intelligence capability and shape it into a product experience that felt clear, connected, and actionable.
That meant understanding the technology, listening to users, aligning with stakeholders, testing assumptions, and designing the system around the decisions people needed to make.
The goal was never to simplify the intelligence. It was to make the intelligence easier to understand—and easier to act on.