UX Project: AI-Powered Benefit Assistant

Company: WEX, Inc.

Location: Portland, ME

Duration: October 2023 - Present

Artificial Intelligence
Visual Design
Product Operating Model
User Experience Design
Cross-Functional Team


Overview

Project Summary & Role

The Benefit Assistant is an AI-driven tool designed to help employees better understand, select, and utilize their benefits. It simplifies complex medical and insurance terminology, provides personalized information, and educates employees on their benefits. The tool offers a chat-based interface that provides 24/7 access to benefits information, while ensuring employees feel empowered to manage their benefits confidently.

View prototype

My Role

I led the design strategy for this application and set the vision for creating the first release of a chat experience. Currently, I am planning a longer-term vision of using AI within the platform in more innovative ways. I directly managed a designer on the project and took an active role as part of the stakeholder leadership team in transitioning to the product operating model. I ensured the design aligned with patterns used across other assistant tools within the organization and provided ongoing feedback on the designs. Our team collaborates frequently with product and tech teams to ensure alignment and smooth execution. We are continuously learning from consumers, identifying their pain points, and iterating based on real-time insights.

Problem & Objectives

Challenge

The biggest challenge is proving employee adoption. If employees do not use the Benefit Assistant, all other goals, such as reducing call center volume, increasing employee satisfaction, and influencing sales, become irrelevant.

Business Impact

Successfully driving adoption will support reduced HR workload, higher engagement, and increased productivity, which will ultimately lead to better benefits utilization and employee satisfaction.

User Need

Employees need a tool that provides easy access to their benefits information, guides them in understanding complex terminology, and helps them make informed decisions about their benefits.




The Process

Discovery

Discovery & Research

  • Usability Studies: We conducted numerous usability studies to refine the design and placement of the chat interface, ensuring ease of access and use.
  • Journey Mapping: Journey maps were created to visualize the employee experience, highlighting pain points and opportunities for improvement.
  • Product Operating Model: We embraced the product operating model, speaking directly with customers (or proxy customers) on a weekly basis to gain insights and iterate based on experimentation.

Design Strategy & Key Features

  • Text Prompts: Clear, concise text prompts guide users through interactions, making the tool intuitive and easy to use.
  • User Feedback Options: Users can easily provide feedback, helping to improve the system’s accuracy and user experience.
  • Friendly, Professional Tone: A balanced tone that is approachable yet professional ensures users feel comfortable while navigating serious benefits-related content.

Interactive Prototypes & Handoff

Prototypes were created in Figma, with the development handoff streamlined through Dev Mode. Design files were organized in workflow order to facilitate understanding and implementation by developers. The design team worked closely with developers on a daily basis for ongoing UX reviews to ensure alignment and smooth execution.

Results & Key Findings

The Benefit Assistant is close to a pilot launch, with early-stage metrics being tracked. We expect to evaluate adoption rates, user feedback, and efficiency improvements once the tool is live. Early internal testing and usability studies have shown promising signs of positive employee engagement.

Metrics to Track Post-Launch

  • Adoption Rates: Tracking employee usage and engagement with the tool.
  • User Feedback: Gathering insights from employees on their experience using the Benefit Assistant.
  • Efficiency Metrics: Measuring the reduction in HR workload and improvements in employee productivity.

Lessons Learned

  • AI Adoption: AI is still a new concept for many users, with some hesitant to share their data. It’s important to balance personalization with transparency and trust.
  • Value of Direct Consumer Insights: Implementing the product operating model and speaking directly with users has been invaluable, providing insights we would not have gathered otherwise.