top of page
24-hour pharmacy sign

Voice UX Research: Real-Customer Testing of a Pharmacy IVR
(Client confidential - project anonymized under NDA)

📌 TL;DR

I led remote moderated usability testing of a conversational IVR system to evaluate prescription refill flows. We achieved a 93% task success rate and 4.8/5 user satisfaction, with findings that informed key design improvements.

🧭 Project Overview

​

Type: Usability testing - Voice/IVR

My Role: Lead UX Researcher / Moderator

Methodology: Moderated remote usability testing

Tools: Hotjar, Simulated voice interface (WoZ), CSAT Survey

Participants: 7 real customers, U.S.-based

​

Context

Users navigating pharmacy phone systems often struggle with unclear terminology, confusing pathways, or overly rigid menu structures. These breakdowns create frustration, long call times, and misrouting.

🎯 Problem Statement + Research Goals

​

Healthcare callers frequently encounter barriers when using automated systems. 

Examples include: 

  • Difficulty understanding menu options

  • Confusion about terminology
  • Navigational dead-ends
  • Misrouted calls due to unclear phrasing
     

Goal: 

Evaluate the existing call flow to uncover usability issues and identify clear, actionable improvements to reduce confusion and streamline user pathways. 

👥 Participants

​

  • 7 participants, recruited via HotJar, from diverse U.S. regions

  • Ages 30-60, a mix of genders and educational levels

  • All were current customers of the pharmacy service

  • Most preferred using digital tools (app/website) but were comfortable calling in when needed.

🔍 Methodology

​

  • Wizard-of-Oz (WoZ) simulation: Participants interacted with a simulated IVR system via phone. Responses were controlled to mimic real system behavior without introducing tech errors.

​

  • Task flows included:

    • Refilling prescriptions (single & multiple)

    • Checking the status of prescriptions

    • Navigating pickup time splits

    • Entering Rx# vs. DOB for authentication

​

  • Tools and Activities:

    • Introductory survey

    • Real-time observation

    • Post-task CSAT (1–5 scale)

    • Qualitative feedback and utterance analysis

 

🧠 Key Findings

​

What Worked Well
✔️ High success rate (93%) and very strong satisfaction (avg CSAT 4.8/5)
✔️ Users found the system clear, quick, and easy to understand
✔️ Voice sounded natural and human-like
✔️ DOB authentication was preferred over entering Rx numbers
✔️ Clear naming of medications helped users feel confident

​

Opportunities for Improvement
⚠️ Users wanted to hear more specific pickup times and price info
⚠️ Lack of transparency around backend processing or delays
⚠️ Some wanted to refill selected prescriptions, not all at once
⚠️ Compared to the mobile app, the IVR felt less detailed in terms of real-time status

 

💡 Recommendations

​

  • Include more specific and actionable pickup timing

  • Surface medication costs, where available, or provide clearer messaging if unavailable

  • Allow users to select individual prescriptions for refills

  • Consider integrating more real-time backend status messaging into the IVR

  • Explore how to match or supplement app functionality within a voice interface

 

📈 Outcomes

​

  • Generated a prioritized list of product improvements

  • Partnered with internal teams to plan and support changes

  • Set the foundation for a follow-up usability test after implementation (Second test ran August 2025, results pending)

 

🧪 What I Brought to the Table

​

  • Designed the end-to-end research plan and script

  • Facilitated usability sessions and led observation synthesis

  • Conducted a detailed analysis of both quantitative (CSAT) and qualitative data

  • Created actionable insights to guide product iteration

  • Advocated for user needs across voice UX and pharmacy flows

 

“I liked that the system used the actual prescription names - it make it feel more human"

*Quotes anonymized for confidentiality

🔒 NDA Compliance Note

Due to a Non-Disclosure Agreement (NDA), client identity and any proprietary content or visuals have been omitted. All findings and insights have been anonymized and represent my contributions.

 

bottom of page