Enhancing Voice AI Experience in Insurance Customer Surveys

Project Overview

Developed UX/UI solutions for a consumer-facing survey app used by insurance companies to securely collect claim-related voice and photo data from customers. The app’s AI-driven voice analytics verified identities as part of the claims process.

Industry Type: Voice Analytics/ AI 

Company Size: 11-50

My role: UX & UI Designer

Project duration: 3.5 weeks

KPI: Achieve a task completion rate of more than 80% (baseline<70%)

Screenshots of a mobile app interface for a questionnaire process, including an audio check, welcome screen, and submission confirmation.

Users struggled to complete voice-enabled questionnaires due to poor sound capture from their smartphones. Optimal voice capture required users to hold phones unnaturally close (6 inches from face), versus typical comfortable distance (9-12 inches), leading to frustration and low task completion rates.

Challenge

User Context

  • Participants complete surveys on personal phones at home, office, or indoor locations, without supervision.

  • Use cases span insurance claimants, job applicants screened for sensitive IPs, and travelers applying for insurance in natural disasters.

Key Opportunities

  1. Educate users on proper phone positioning for optimal voice capture.

  2. Simplify interactions to minimize user confusion and errors.

  3. Deliver clear, unambiguous instructions to improve task completion.

Design & Iterations

Iteration 1: Educational Walkthrough

  • Created mid-fidelity walkthrough screens illustrating correct phone angle and distance before the questionnaire start.

  • Provided visual context explaining why errors occur to improve user understanding and reduce friction.

  • Implemented as a fast, low-cost interim solution aligned with limited resources and timeline.

Instruction screen on a smartphone showing how to hold the device for face capture, with illustrations indicating holding it within 6 inches, and making sure the face fits inside the inner circle on the screen. 
Mid-fi walkthrough screen

Usability Testing Insights

  • Sound quality varied by device and OS, limiting consistency.

  • Lower digital literacy correlated with reduced task success.

  • Multiple pre-survey steps caused uncertainty around task length and progress.

A smartphone displaying a blank white screen with a small circular loading icon at the bottom center.

Questionnaire application before revamp

Iteration 2: Application Revamp

  • Introduced a real-time visual cue: a framing box that alerts users when the phone moves too far from the face, guiding optimal positioning.

  • Added a system status bar to indicate remaining steps, enhancing perceived progress and reducing uncertainty.

  • Revised call-to-action buttons with descriptive labels to clarify next steps.

  • Incorporated stakeholder feedback by adding a highlight tint to the framing box for better visibility.

Empty smartphone screen with a white background.

Redesigned questionnaire application

  • Improved the video check interface based on ongoing feedback to further streamline user experience (details pending production).

Iteration 3: Video Check Screen Update

Smartphone screen showing a video check for face recognition, with a photo of a smiling man with glasses inside an oval frame and a message to hold the device closer.

Final video check screen

Next Steps

Analyze post-launch data on learnability and first-attempt completion rates to validate design effectiveness and inform future enhancements.

Thank you for reading!