Hi, Mike.

If I’m understanding your question correctly, it seems like creating audiences is the answer.

You can create audiences here: Settings > Data > Audiences. Just search “Audience” in Settings. It sounds like you’ll want three audiences, one for each product/app the user is chatting for help from. We’ll just call these Audiences A, B, and C.

Then, over in the Knowledge Hub > Sources, click on the website containing the knowledge base for Audience A’s product (this may look different if you use Intercom for your Help Centers). Next, click the checkbox next to Title to select every article in this knowledge base, then click More Actions > Change Fin audience, and select Audience A. Finally, repeat for B and C.
 

Setting the audiences this way should work for you.

Hi ​@Mike Simmons 

Paul here from the friendly support engineering team!

Fin doesn’t currently use conversation tags to filter Help Center content, but you can guide it in a few effective ways:

1. Update Fin Guidance
Add specific instructions like:

“If the conversation is tagged Product A, only use articles from Product A’s Help Center.”

This helps Fin bias answers based on context, though it doesn’t strictly filter content.

2. Use Custom Attributes Instead of Tags
Pass a product_name attribute to Fin via Messenger, and refer to it in Guidance. Fin responds better to known attributes than post-chat tags.

Let me know if you want help writing the Guidance or setting up product-specific routing!