A user survey collects direct feedback about a chatbot experience, revealing frustrations and expectations that conversation metrics and logs may only suggest indirectly. Keep it brief because longer surveys reduce completion rates. section 3.3.2 section 3.4.1
From complaint to improvement target
A user reports that vaccine-eligibility answers ignore different states, while another finds appointment scheduling confusing and a third waits nearly five minutes for availability. Categorizing these responses turns frustration into concrete problems: missing context, unclear guidance, and excessive response time. section 3.4.1
Use surveys as part of an improvement loop:
- Ask about ease of use, helpfulness, and overall satisfaction after the interaction. section 3.3.2
- Combine responses with conversation transcripts to distinguish reported frustration from inferred failure. section 3.4.1
- Look for repeated problems before prioritizing solutions; isolated responses may not represent the broader user population. section 3.4.1
- Encourage participation with a small incentive or a genuine thank-you, while recognizing that response rates are often low. section 3.3.2 section 3.4.1
