Guide
Manually sorting through the overwhelming volume of feedback is neither efficient nor pleasant. AI changes how we handle, analyze, and respond to user feedback. In particular, it comes handy for generating conversation summaries. Such short summaries are helpful for other teammates to have a glance at topics that were discussed.
In this guide, we'll explain how to configure AI in Automations to complete this task in Fibery automatically.
In the Product Team template, the Automations that rely on AI are described but disabled by default. Once you've set your OpenAI API key, you can enable them and tweak the prompts to your liking. To see an example of generating feedback summary, go to Feedback Space and proceed to Automations in the Interview Database.
How it works
To automatically generate summary for a conversation with a customer, use AI in Automations.
Here is the prompt for AI:
[ai temperature=0.5 maxTokens=300]
"""
{{Messages}}
"""
Summarise the text above
[/ai]
Example
Let's say we use Intercom for customer success, and we want to automatically summarize the chat when its closed.
So we set an Intercom Integration, and now we have a Conversation Database were all chats are stored.
This how the Automation Rule will look like:
To control the length of the summary, adjust maxTokens parameter.
We recommend storing the Summary in a separate Rich Text Field to easily distinguish it from the other notes or conversation itself.
How to visualize summaries
Dynamic Feed View is handy when you need to review text of many Entities, so we recommend using it.
This is how it Intercom integration Feed from the example may look like:
The biggest limitation here is volume. Sometimes users prefer to reopen a conversation with a new question instead of creating a new conversation. If this becomes the preferred way of communication, summarizing effectively becomes less effective over time.
See also