How to use thumb up and thumb down to improve the Virtual Agent response quality?

[Lisa71]

How does ServiceNow use thumb up and thumb down to improve the Virtual Agent response quality? How shall we track the thumb up and thumb down data and use them to improve the Virtual Agent response quality? Thanks.

[Dr Atul G- LNG]

Hi @Lisa71 

https://www.servicenow.com/docs/bundle/zurich-conversational-interfaces/page/administer/now-assist-i...

Each Virtual Agent response includes a feedback icons panel. The feedback icons panel appears on the latest Virtual Agent response and whenever you hover over any previous Virtual Agent response. You can indicate if the response was helpful by selecting the like thumbs up icon (

). If the response wasn't helpful, select the dislike thumbs down icon (

). This feedback is used to train the LLM model and improve responses over time. Depending on the context of the response, an additional go to search results icon (

) may appear in the feedback icons panel. This icon appears alongside synthesized responses in Virtual Agent, clarifying questions in Virtual Agent, and fallback topics whenever a synthesized response is unavailable but there are regular search results available. When the full-page experience is on and you select the go to search results icon (

), you’re redirected to the Search tab. After redirecting to the Search tab, a search query using the last five chat utterances you entered begins.

https://www.servicenow.com/community/virtual-agent-nlu-forum/how-to-check-and-display-the-thumbs-up-...