AI Powered Help Center

User Problem:
How can we incentivize users to resolve issues using non-human support methods (e.g., AI chat, help center articles)?

Key Challenge:
AI-powered chat costs under $1 per session versus $3-4 for human support. Eventbrite sought to deploy AI for self-serve issues resolvable through existing articles. However, users consistently bypassed alternative channels in favor of direct human support, even for self-serve problems. Additionally, Eventbrite's intended AI use cases exceeded the technology's current capabilities.

Method:
1. Internal Research Audit: Consolidated fragmented and ambiguous data across teams
2. Competitive Analysis: Updated 2-year-old analysis to examine how competitors integrated AI solutions (conceptually and design wise)
3. Usability Testing with In-depth Interviews: Validated design concepts and assessed incentivization effectiveness and user flow

Findings:
- Two factors influenced human support preference: age and perception of speed
- Users preferred varied resolution methods but defaulted to human support, perceived as fastest or most reliable
- Tools misaligned with user expectations eroded brand trust
- Competitive insights informed design improvements: introduced new resolution methods, streamlined help center interface, and reduced cognitive load through improved navigation

Impact:
Research findings led to three outcomes: (1) revived the help center redesign initially launched in January 2025 and paused until research completion in October 2025; (2) reversed Eventbrite's proposed human support prioritization strategy to protect client relationships; (3) delivered a redesigned platform with measurably reduced time-on-task and increased CSAT scores (+0.7 on Likert scale).