Plan requirement
| Subscription | Any plan |
| Access | Admin |
Patterns in what customers say, across many tickets. You install it and it analyses on its own: there is no period or scope to set. What decides whether it works for you is which tickets it can see and how many of them there are.
What it needs before it says anything
| Which tickets | Solved or Closed only. |
| How many | About 50 solved or closed tickets matching a category's filters and time period, per category. |
| First results | Within 24 hours of installing. |
| After that | Patterns refresh automatically every 12 hours. |
You need to be an admin to install and use it.
What it cannot see
Phone and voice, social media, and help centre article comments are not analysed, and AI agent conversations are excluded from pattern generation entirely.
That matters for the conclusion you draw. If a third of your contact arrives by phone, the patterns describe the other two thirds, and a theme that looks like it is fading may have moved to a channel the app does not read.
What a pattern shows you
Each one carries a title and summary, root causes, suggested actions, customer feedback, the relevant tickets, and a generated date with a confidence indication. You can sort and pin them, and ask an AI assistant follow-up questions about a pattern.
Open the relevant tickets anyway. A pattern is a label on a pile, and reading five tickets behind it frequently changes the interpretation. The generated date and confidence are there for the same reason: they tell you how much weight the label can carry.
Acting on the label without reading the tickets is how a support team solves a problem nobody has.
Where it earns its place
After a release, to see what changed in what people mention. Before a planning cycle, to argue from what customers said rather than from what the loudest colleague remembers. And on a queue nobody has time to read through, which is most of them.
Comments
0 comments
Article is closed for comments.