Best practices when using FAQ Plus

• Knowledge base quality: A good bot experience for your end users is critically depended upon the quality of your knowledge base. Please refer to this links on helpful best practices on how to set up useful knowledge bases.

  • https://docs.microsoft.com/en-us/azure/cognitive-services/qnamaker/concepts/best-practices
  • https://github.com/MicrosoftDocs/azure-docs/blob/master/articles/cognitive-services/QnAMaker/Concepts/best-practices.md

Tip: If you don't have the complete knowledge base ready, here is a Teams Frequently Asked Questions Link that works well with QnA Maker.

• Managing user expectations: The bot always keeps the end user apprised of the status of their requests. For example, the bot will notify end user any time the status changes (e.g. from unassigned to assigned). You should also leverage the help tab text to define and set expectations for requestors. For example, if you goal is to answer all queries within 24 hours, mention that within the help tab.

• Setting your Expert team up for success: Similar to knowledge base quality, the success of this app in your organization also depends hugely on the team of specialists/Subject Matter Experts you have. You would want to make sure they can reap the full benefits of this app in helping end users. Following are the items you should consider:

  • Involve them in the installation process, show them how the full app works - there are three big components - end user bot; bot notifications in the general channel of the team; Messaging extension that enables experts to quickly sift through assigned/unassigned request.
  • Set up expectations with end users and empower the specialists to deliver those expectations. For example, ensure that all queries do get resolved within 24 hours or the end user is apprised of what's taking longer than usual. This will build confidence and you'd have returning users.
  • Leverage the 'Help Tab'. Use this tab to document process you'd like your end users to follow. For example, giving them a lay of the land on what kind of questions they can ask and under what conditions they should escalate to asking an expert are good candidates for such content.