Skip to main content
  • There is currently not an easy/intuitive way for an agent to know before leaving a comment if the question they want to ask has already been answered, or if the suggestion they want to make has already been proposed.
    • The comment icon at the top of the Guru card only shows unresolved comments, and it’s not really made obvious where else someone would look for resolved comments. The “resolved comments” tab also doesn’t show a number indicating how many resolved comments there are the same way that the “cards that link to this card” tab does, so if there’s no number displayed, a user may just assume this means that there are none.
      • (I know that the tab simply doesn’t display at all if there are no resolved comments, but someone unfamiliar with the ins and outs of Guru/who doesn’t look at card history very often wouldn’t necessarily automatically assume this to be the case)
  • Sometimes we get multiple comments on a Guru card that have the same answer but don’t make sense to add into the card itself - things along the lines of “why is our policy ABC instead of XYZ?” that would just disrupt the flow/not fit the purpose of the card.  Suggestions:
    • Make the comment history of a card more obvious and easy to access - maybe as an option that lives in the same area as the Table of Contents, and similarly can be popped in and out. Only one click required, rather than having to navigate multiple menus.
    • If a user is commenting on a specific section of a card, surface comments that have previously been left on that section. If the user sees a similar comment has already been made & resolved, no need to post theirs.
      • Or a feature similar to how when I’m writing the title for a post in this forum, a “related topics” pop-up suggests existing posts that might be similar to the one I’m about to make. If they start typing “What should I do if a customer says XYZ” then any similar resolved comments are surfaced.
    • Integrate Knowledge Agents into the comment feature.
      • If a comment is a policy/process question, the AI could attempt to generate an answer and ask “Does this answer your question?”
      • If the agent says Yes, the comment is discarded. If they say No, the comment gets posted.
      • Our Knowledge Agent adoption is pretty low, with a lot of agents using comments to ask questions that the Knowledge Agent could answer, so this could be a way to drive more adoption of Knowledge Agents.
Be the first to reply!
Terms & ConditionsCookie settings