Knowledge Base
Was This Helpful? Feedback
Every published article shows a thumbs-up / thumbs-down prompt - the responses are tracked per article and visible in the help center analytics.
Every published help article shows a “Was this helpful?” prompt at the bottom. Customers can give a thumbs up (helpful) or thumbs down (not helpful).
This gives you a direct signal from customers about which articles are actually useful and which need work.
What gets tracked
For each article, Convot tracks:
- Helpful count - total thumbs-up responses, lifetime.
- Not helpful count - total thumbs-down responses, lifetime.
These are shown in the help center analytics dashboard alongside view data.
Where to find the feedback data
Go to Reports - Help center / portal analytics. The feedback section shows:
- Total helpful vs not-helpful across your whole knowledge base.
- A list of the lowest-rated articles - those with the most “not helpful” votes relative to “helpful” - so you know exactly which articles customers find lacking.
See Help center analytics for the full dashboard.
Acting on feedback
When an article has a high “not helpful” rate, it typically means one of:
- The article does not answer the question customers are actually asking (the title promises something the content does not deliver).
- The steps are outdated or incomplete.
- The article covers the topic but misses a common variant of the question.
Check the article alongside the AI agent’s knowledge gaps report - if the same topic appears in both, it is a high-priority fix.
The AI agent connection
Article feedback also informs the AI agent’s quality signal. When a customer marks a bot answer as unhelpful and the bot was drawing from a specific article, that interaction appears as a “weak answer” gap in the knowledge gaps report - pointing you to the article that needs improvement.
Was this article helpful?
Thanks for your feedback!