Scope AI replies by inbox?

Question

Hi,We have a single workplace with two inboxes, each representing a different product. AI is enabled for both inboxes, and we use shared rules and templates.We’ve noticed that AI-generated replies (Copilot, and I assume Autopilot as well) will sometimes reference information from the other product when composing a response, even though the conversation is clearly in a specific inbox. As a result, the replies sometimes need manual editing to correct product-specific details. For example, the AI may compose a response for one inbox that should only apply to another product’s inbox.It appears that AI responses are not factoring in inbox context when generating replies.Is there a way today to scope AI behavior by inbox? If not, is inbox-level AI scoping something on the roadmap?Thanks!

cat_apac · Answer

Hi there! Cat here from the Front support team 🌻

Right now, Copilot pulls from all your connected knowledge sources when generating replies, so it doesn't have a way to limit its info to just one inbox in a workspace. That means if you’re working across multiple inboxes that share knowledge content, AI replies might occasionally mix in details from another product or inbox (totally understandable to want more separation!).

Copilot can tell when you're switching between different workspaces, but within a single workspace, there isn’t currently an option to restrict AI responses by inbox. This is how things are set up at the moment, but if you have recommendations on how we can better handle this, we’d love to hear your thoughts and ideas in our Product Ideas Portal!

Cheers,
Cat

Sign up

Use your Front credentials

Login to the community

Use your Front credentials

Scanning file for viruses.

This file cannot be downloaded