American Express · 2022 – 2025
Intuitive Servicing Portal
Building a research practice and designing a dozen-plus servicing journeys for a global internal platform.
01
Overview
At American Express, I worked across UX research and UI design to support the development of the Intuitive Servicing Portal (ISP), a new internal platform designed to consolidate tools and workflows for global Customer Care Professionals (CCPs). ISP aimed to replace outdated, fragmented processes with a unified, intuitive interface that reduced inefficiencies and improved the servicing experience for Card Members worldwide.
Prior to ISP, CCPs relied on a legacy PEGA-based system with complex workflows that often required navigating multiple screens, switching between tools and stepping out of the portal to access relevant resources. These challenges extended call handling times, increased the potential for errors and made onboarding new hires more resource-intensive.
My role involved deeply understanding these operational pain points through contextual inquiry, interviews and usability testing, then translating findings into actionable design solutions that aligned with brand standards, accessibility guidelines and regulatory requirements. I partnered with product managers, developers and compliance leads to make sure solutions were not only user-friendly but also feasible within technical and business constraints.
A note on confidentiality. This case study has been abstracted to protect confidential American Express information. Proprietary UI designs and internal materials aren’t shown here, but I’m happy to discuss them in interviews.
02
Defining the research process
When I joined the Capabilities and Servicing team, UX research was still a new discipline. A colleague and I developed a formal research intake process for designers and stakeholders, along with foundational research standards to guide future projects. The framework brought clarity, consistency and efficiency to how research was requested, planned and executed, and it’s the foundation the Membership Rewards work was built on.
03
Research approach
Contextual inquiry & interviews
Shadowed and interviewed CCPs across multiple departments (Membership Rewards, Credit & Fraud, Capabilities) to understand real-world workflows, mental models and departmental differences.
Card sorting
Restructured the information architecture for the Customer Help Center (CHC) tool with Optimal Sort, revealing how different departments organized and accessed information, then standardized navigation across teams.
Usability testing
Designed and moderated sessions with 30+ CCPs, evaluating new ISP designs for clarity, efficiency and ease of navigation, and collecting task success metrics alongside qualitative insights.
Compliance & localization
Led research into regional regulations and language requirements for the UK and Canada, so solutions could adapt to multiple markets without sacrificing usability.
Findings were consolidated into actionable insights and shared through workshops, reports and visual artifacts that kept stakeholders aligned and guided design priorities.
04
Synthesized artifacts
Observational studies were synthesized into journey maps that made pain points tangible for stakeholders.
05
Journeys I designed
I designed and delivered more than a dozen high-impact journeys across ISP.
- Pay with Points history & reversal
- Gift card redemption, history & reversal
- Statement credit / redeem for deposit
- Points transfer redemption
- Cover Your Charges redemption, history & reversal
- UCID (manual authentication / secure push), adaptable for multiple markets
- Charge verification, with guided workflows that reduce resolution time and errors
- Traveler’s Cheque replacement
- Card replacement
- Cancel card, with tailored messaging for different scenarios
- Claim work basket for queued service requests
06
Design solutions
- Embedded knowledge resources: integrated relevant CHC article content directly into the ISP workflow, reducing “step-out” actions and keeping CCPs in a single environment.
- Streamlined navigation: reduced multi-screen workflows into consolidated task flows, minimizing tool-switching.
- Localized experiences: designed flexible interfaces for multiple languages, market-specific content and varying compliance requirements.
- Error prevention: added inline validation, contextual guidance and clear feedback states to reduce transaction errors and rework.
- Consistent UI: kept a cohesive look and feel with shared design libraries, standardized components and accessibility best practices.
- Scalable framework: created solutions that could expand across journeys and departments without major structural redesigns.
07
Outcomes
Reduced call handling time
Embedding key CHC content into ISP workflows cut average handling time and eliminated unnecessary step-outs.
More consistent operations
Standardized navigation and task flows reduced variability in how CCPs completed similar tasks.
Faster onboarding
Simpler workflows and clearer UI patterns lowered the learning curve for new hires.
Stronger stakeholder alignment
A repeatable research intake and reporting process kept stakeholders engaged throughout each phase.
Global readiness
Solutions adapted to multiple markets, languages and regulations without sacrificing usability.
08
Key learnings
- Department-specific needs matter. Early card sorting showed that departmental context strongly influenced how CCPs navigated tools, which led to more universal but flexible navigation.
- Integrating research early yields stronger designs. Embedding research from project inception meant solutions addressed root causes rather than symptoms.
- Cross-functional collaboration is critical. Partnering closely with product, engineering and compliance balanced user needs with technical and regulatory realities.
- Scalable design benefits the whole system. Reusable components and workflows made it easier to expand ISP without redesigning from scratch.
- Continuous feedback loops drive improvement. Iterative testing and synthesis kept designs grounded in real user feedback rather than assumptions.