Click on the Knowledge Gap feature to get started.
You will be redirected to the Knowledge Gap page.
Step 2: Connect Help Desk
Click on Connect to link your help desk.
Connection options will be presented.
Step 3: Select Knowledge Base
Choose the knowledge base for which you want to run the Knowledge Gap.
Step 4: Proceed to Next
Click Next to continue.
Step 5: Select Help Desk
Select the help desk you are currently using.
For this instance, we will choose Groove.
Step 6: Enter API Token
Enter the API token for the selected help desk.
Step 7: Select Mailbox
Select the mailbox from which tickets will be used to identify gaps.
We will select Groove Support.
Step 8: Start Gap Search
Click on Find Gaps to start identifying gaps.
It will first analyze 5,100 support tickets from the last three months. Then, at the end of every week, Knowledge Gap will automatically run and analyze 300 tickets from that week.
Step 9: Enter Website
Enter your website associated with the products.
We ask for your website so we can generate a company profile automatically (which you can edit later). This profile helps our AI understand what your company does, what your knowledge base is about, and identify content gaps that are most relevant to your business.
Step 10: Process Tickets
It will automatically begin by processing the first 100 tickets and identifying gaps. Once that’s complete, it will proceed to analyze all 5,000 tickets. You can see the real-time progress too.
Step 11: View Gap
Monitor the gaps being generated from the analyzed tickets.
We automatically suggest fixes for the first five gaps so you can clearly understand what’s happening. That’s why the first five are marked as “In Progress” while the rest haven’t started yet.
From here, you can review the priority, open each gap, let the AI suggest a fix, whether that means creating a new article or updating an existing one and then publish the content.
FAQs
Why does it analyze fewer tickets than expected?
This can happen if you don’t have enough tickets for that time period, or if some of the tickets we analyzed didn’t contain sufficient information. In those cases, we reject the ticket after analysis, and it isn’t counted in the “analyzed tickets” metric.
