From the Content page, you can create and manage snippets and enable them for use with Fin AI. Snippets are small sources of information which aren’t publicly available to customers and only used by Fin to improve answer quality and coverage.
Create a snippet
To create a new snippet, go to Train > Sources and click on + New content then select Snippet.
Add content
Now you can create your snippet by entering a title and body in text format.
You can highlight text to format it such as adding headers or links.
In order to get the best results, we recommend following these best practices.
Save it
When you’ve finished writing the snippet, or you want to navigate away from the editor, you can click Save in the top right of the screen to save your changes.
Note:
Your snippet will also be auto-saved every few seconds while you're working in the editor. If your changes have been autosaved, the Save button will be grayed out and you can simply click on Cancel to exit the editor.
There is no “Save as draft” option for snippets. When snippets are saved, this is the information that becomes available to customers through Fin AI (if enabled).
Once saved, snippets will usually be available for Fin to use within 10 minutes.
Manage snippets
Managing a snippet is all done from the snippet editor. Go to Train > Sources and click on Snippets, under the More content sources heading, then click on the snippet you want to manage.
Configure settings
When you create a snippet, there's a "Details" panel on the right which contains:
Data: View the content type, language, creation date, created by, last update, and the teammate who last updated it.
Fin settings: To enable/disable for Fin AI. When enabled, the content becomes available to customers through Fin AI, respectively.
Audience: Target your content at a specific audience, so your customers only see the content that is relevant for them.
Reports: View you reports for a deeper understanding of how your customers interact with your content.
Folder: Add the snippet to a relevant folder in your content page to keep content organized.
Make changes to the content
To edit an existing snippet, go to your Content page, click on Snippets, then select the snippet you want to make changes to and click Edit.
When you’ve finished making changes, or you want to navigate away from the editor, click Save.
Your snippet will also be auto-saved every few seconds while you're working in the editor. If your changes have been autosaved, the Save button will be grayed out and you can simply click on Cancel to exit the editor.
Make it available to Fin
To make a snippet available to Fin AI, go to Train > Sources > Snippets and open the snippet you want to enable.
From the "Details" panel, scroll down to “Fin settings” and toggle on Fin AI Agent - This setting will make the snippet available for Fin AI to use when responding to customers (it will respect any audience rules).
Note: The language of the snippet must be one of the supported Messenger languages otherwise the "Available to AI Agent" options will be grayed out.
Make it available to a specific audience
If your snippet is only relevant for a specific subset of customers, you can use audience filters to make it visible to certain people.
First, you’ll need to create and define the audience you want to target.
Then go to your Content page and open the snippet. From the "Details" panel scroll down to “Audience” and use the dropdown to select one of your pre-defined audiences.
Note:
The default audience for snippets is “Everyone”.
Fin AI will also respect any audience you apply to a snippet and only use this snippet to answer customer questions if they match the audience rules.
Fin AI Agent answer behavior
When a snippet is used as a content source in a Fin AI Agent answer, the customer is not shown the link to the snippet. Instead, they will see a message saying: AI answers are generated based on both public and private sources provided by [your company name].
Best practices for creating effective snippets for Fin
Snippets are a powerful tool to ensure AI agents, like Fin, can provide accurate and helpful responses to customer queries. Follow these best practices to optimize your snippets for clarity and functionality:
1. Avoid ambiguity
Clear Titles: Ensure the snippet title explicitly defines the product, feature, or service it pertains to. Avoid vague or overly broad terms.
Specific Content: The body of the snippet should leave no room for misinterpretation. Clearly explain the topic to eliminate confusion.
2. Restate questions in the body
To enhance Fin’s ability to match customer questions with snippets, restate the content from the title in the snippet’s body. For example:
Title: "How to reset your password"
Body: "If you’re looking for instructions on how to reset your password..."
3. Use headers for structure
For longer snippets (200+ words), divide the content into distinct sections using headers (H1 or H2). Each header should succinctly summarize the information that follows.
Example:
Header: "Steps to Reset Your Password"
Content: "Follow these steps to reset your password..."
4. Utilize lists
Break down detailed information into lists instead of lengthy paragraphs. Fin is better at processing and delivering structured, concise points.
Example:
Instead of: "To reset your password, go to the login page, click on 'Forgot Password', enter your email address, and follow the instructions sent to your inbox."
Use:
Go to the login page.
Click on "Forgot Password."
Enter your email address.
Follow the instructions sent to your inbox.
5. Optimize snippet titles with variations
Include multiple variations of the question or topic in the snippet title. Separate these variations using punctuation such as slashes (/), question marks (?), or commas (,).
Example:
Title: "How to reset password / Forgot password / Password recovery instructions"
Since snippet titles aren’t visible to customers, you can make them as descriptive and lengthy as needed to cover various phrasings customers might use.
By applying these best practices, you’ll ensure your snippets are both effective and AI-friendly, allowing Fin to deliver accurate and contextually relevant answers to your customers.