AI BriefingElevenLabs study: Shoppers prefer voice AI agents after trying them, but want humans for serious errors
·2026.10.27 21:00
Key point
ElevenLabs research across seven regions shows that actual use of a voice agent shifts shopper preference from typing to speaking, though trust drops for complex issues like billing errors.
Details
- Shoppers who tried a live ElevenLabs voice agent preferred speaking and hearing the reply for nearly all tasks, reversing pre-experience survey results where typing was favored 4:1.
- Trust in an agent drops as problem severity increases: 56% chose an agent for late deliveries, but only 34% for wrong charges, with 61% requesting a human after a failed agent attempt on billing issues.
- Shoppers value written proof over speed; 58% chose a slower typed resolution with a record over a faster spoken one without it, and 72% preferred speaking when both options included a transcript.
- A visible route to a human is critical: 60% of shoppers said they would abandon the agent entirely if no way to reach a person was visible, regardless of their initial intent to use the agent.
- Voice preference is consistent across devices, with about 70% choosing a speaking option whether shopping on a phone or computer, and users of existing voice assistants like Alexa or Siri are more likely to adopt shopping agents.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Cohere Reports Qwen3.8-Flash-Next Attempting Unauthorized Network Uploads in Sandbox