Led the end-to-end design of a new workflow for evaluating conversational voice agents- from product definition and stakeholder alignment to interaction design and deployment readiness.
UX Design
AI
Featured
The Opportunity
Voice agents and assistants were becoming a major capability across the industry and on our platform, but unlike chat agents and workflows, there was no way to reliably evaluate their quality before deployment.
Businesses across the world are starting to rely more on voice agent for customer support and automated ticket handling. The market is a rapidly growing with multiple early players and competition from major players is less.
The case for voice agents is also stronger with the costs saved on every solved case.
With increasing AI call volumes, it became necessary to automate the evaluation for voice agents to test the scale and breadth of the possible calls to the agent and the quality of the responses and executions.
The Problem
Why simply using the same chat agent evaluation platform wasn't enough.
Chat agent/workflow | Voice agent/assistants |
|---|---|
Text input | Speech input |
Smaller conversations, shorter queries and responses | Longer multi-turn conversations, longer queries |
Instant simulation | Long simulation time |
Only intent | Intent, tone, interruptions, latency, environment |
Static prompts | Dynamic scenarios |
Voice modality carries a lot more complications and simply copying the existing paradigm wouldn't work.
The Process
The Outcome
