Skip to main content
Fin AI Agent unresolved questions

How to find questions that are unresolved by Fin AI Agent and pinpoint areas for improvement in your content or Fin setup.

Beth-Ann Sher avatar
Written by Beth-Ann Sher
Updated over a month ago

Quickly gain insight into the questions that are unresolved by Fin AI Agent and pinpoint areas for improvement in your content or Fin setup.


Get started

Go to Fin AI Agent > Overview and click on the Content gaps tab.

You'll find a list of unresolved questions by Fin AI Agent. These are automatically grouped by topic areas you can focus on to increase the number of resolved conversations.

If you don’t see a list of questions, we are still gathering insights. Once you have enough similar unresolved questions, your insights will be displayed. There isn't a minimum number of conversations required, it'll depend on each workspace. Consider increasing the number of conversations Fin AI Agent gets involved in (Involvement Rate) in order to see insights sooner.

Each grouping is a cluster of similar questions asked by your users that were unresolved — that is, Fin AI Agent either didn’t have an answer for the question, the user asked to be routed to the team, or the conversation was abandoned. The question group naming is created by AI and summarizes the questions in that particular group.

You can sort the list of unresolved question groups by:

  • Language: language of the questions and conversations in the question group (this can be filtered).

  • Volume: total number of conversations in a question group.

  • Abandoned: number of conversations where Fin sent an answer, the customer gave negative feedback, and left the conversation.

  • Routed to team: number of conversations where Fin gave an answer, but the customer asked to talk to the team.

This also includes data from Custom Answers if you have these enabled.

Deep dive into conversations

Click into each question group to see a list of associated conversations open up in the side panel. Click on a specific question to open a new tab and look at that conversation in more detail and see the teammate's response. Their response will help you identify what information is currently missing from your content.

Focus on the most specific question groups, e.g. "Project duplication issues" would help highlight specific issues customers are having with this feature, vs "Task management" which could be a broad area and involve more manual investigation to drill down into conversations and find content gaps.

Improving your content

Easily find related content Fin AI Agent has used in attempts to answer these questions and update existing content or add new content to improve how Fin answers this question in the future.

To see related content Fin AI Agent has used, click on a question group to open the side panel, then click on the Related content tab. This will provide a list of sources Fin AI Agent used to try answer the questions in that group, including:

  • Content type (article, snippet, etc.)

  • Frequency (number of conversations in which this source was used)

  • Last edited at (when this content was last updated by a teammate on your workspace)

Clicking on a content source will open the content in a new tab for you to edit.

From the side panel you can also click on the Add content button to create a new piece of content that answers the unresolved question.


How the data is generated

We analyze up to 10,000 conversations per week. If you have more conversations on a weekly basis, we will take a sample of your conversations to perform this analysis.

For example:

  • If you have 8,000 conversations in week 1, this will analyze 8,000 conversations. If in week 2, you have 12,000 conversations, the analysis will cover 10,000 conversations out of a total of 20,000. If in week 3 you have 5,000 conversations, the analysis will cover 10,000 out of 25,000 conversations, and so on.

  • If you have 1,000 conversations in week 1, and 500 conversations in week 2, and 3,000 conversations in week 3, and 500 conversations in week 4, at the end of week 4, an analysis will be done on all 5,000 conversations.

Data refresh frequency

The reporting data uses data for the conversations of the past 30 days, and gets refreshed and fully regenerated every week. When there is a new set of question groups for the workspace, and you haven't seen them, an Updated stamp on the Content gaps tab will be shown.

Note:

  • If you haven’t set Fin AI Agent live or Fin hasn't been involved in any conversations yet, you'll see an empty state on the Content gaps tab.

  • Once Fin is involved in conversations, you'll start seeing data here after the first week.

  • You'll need a minimum of 10 unresolved Fin conversations a month for insights to be generated.


💡Tip

Need more help? Get support from our Community Forum
Find answers and get help from Intercom Support and Community Experts


Did this answer your question?