Inside the six AI agents built into Inbxer

AI in Inbxer isn't a chatbot bolted onto the sidebar. It's six narrow agents, each doing one job on a thread, and every one of them stops short of sending anything without you.

The six agents

• Thread summary agent — condenses a long back-and-forth into a few lines an agent can read before picking up the thread.
• Thread urgency agent — flags threads that look like they're heading toward an SLA breach or need eyes now, not at the end of the queue.
• Tone rules agent — checks a draft reply against the tone guidelines you've set before it ever reaches a customer.
• Thread labeling agent — applies labels automatically based on what the thread is actually about, not just its subject line.
• Support reply agent — drafts a full reply grounded in the thread's history, ready to edit or send.
• Thread title agent — writes a clean, readable title for threads that arrive without a useful subject line.

Draft, don't autopilot

Every agent produces a suggestion, not an action. A drafted reply sits in the composer until someone clicks send. A handoff detector watches for the cases an agent shouldn't touch — anything that reads as legal, security, or simply outside what the model should be guessing at — and routes those straight to a human instead of drafting anything.

Where the data goes

Where you enable AI drafting, the request goes to the model provider you select. It isn't retained for training by Inbxer. Your selected provider's own data terms apply. If you'd rather not send anything to a third-party model, the agents are optional and the rest of the platform works without them.