The suggest rate is the percentage of questions Answer Bot was active for, and has sent suggestions for. If the suggest rate is low, or lower than expected, it can indicate that your Answer Bot labels are misconfigured, or that you receive a larger-than-usual number of questions from unsupported languages.
This article contains the following sections:
Improving your suggest rate for emails
Check your Answer Bot notification body: Your Answer Bot triggers require the Answer Bot placeholders {{answer_bot.article_list}}
or {{answer_bot.first_article_body}}.
If these placeholders are not present in the body of the email notification then the trigger will still fire; however, no articles will ever be suggested which may cause the suggest rate to reflect inaccurately. Learn more about creating and converting triggers.
Check your Answer Bot labels: If you're using labels with Answer Bot this may be the cause of a lower suggest rate. Labels restrict which articles Answer Bot can use when searching for an article - if it can't find a relevant article in the restricted set, it simply won't suggest anything, causing the suggest rate to decrease. Learn more about using labels with Answer Bot.
Improving your suggest rate for web forms
Check your Answer Bot labels: As with Answer Bot for email, if you're using labels with Answer Bot this may be the cause of a lower suggest rate. Labels restrict which articles Answer Bot can use when searching for an article - if it can't find a relevant article in the restricted set, it simply won't suggest anything, causing the suggest rate to decrease. Learn more about using labels with Answer Bot.
2 Comments
I have a Suggestion Rate of 78.3% in Explore. Is that good or bad? Answers are 808, unsuccessful attempts are 224 and Attempts 1,000. Can you help me to under these numbers?
I have a click-through rate or 12.4%, resolution rate of 1.5% and a rejection rate of 2.2%
Can you help me understand this data for Guide? I do not see any examples of Guide data in Explore. Am I missing it?
Hi Rebecca McMurry,
That is a GREAT question, and the frustrating part is that the answer largely depends on a lot of factors and your use case in general.
I would recommend going through some of your tickets where the bot didn't suggest something, and do a quick analysis of those and ask yourself if there is an article of yours that you think should have been suggested?
To give you a peak behind the curtain a tad, one of the things we do at Zendesk is analyze tickets in this fashion and for tickets that it should have suggested an article and didn't, we try to edit that article so it's more friendly with the customer's language and if there were no articles out there for the bot to suggest then maybe it might be time to write a short article. Answer Bot works best when you feed it a lot of articles.
Long story short, often improving the Bot's suggestion rate and resolution rate involves either making new articles, tweaking existing articles by changing the terminology or updating labels, or maybe updating the trigger that sends the Answer Bot message to make sure it isn't missing certain tickets.
Please let us know if you have any further questions and we'll be glad to help!
Please sign in to leave a comment.