Skip to Content
Torben AI

Example prompts

Ready-made requests for building, diagnosing, and running your support with Torben.

Start with one concrete goal. Include the ticket number, Playbook, or agent name when it matters. Prompts stay editable until you send them.

Build

  • Answer recurring questions. “Create a setup that runs on every new ticket, searches the knowledge base for a matching answer, replies automatically if it finds a confident answer, and assigns the ticket to the support team if nothing relevant is in the knowledge base.”
  • Order tracking. “Handle order tracking questions by looking up the order from the conversation, fetching its current shipping status from the connected integration, and replying with a clear update. If the order is already delivered, resolve the ticket automatically.”
  • Returns and refunds. “Process return and refund requests by looking up the customer's order, checking whether it is within the return window and meets the return policy, and either starting the refund or escalating to a person with a summary of why it needs manual review.”
  • Address change. “Handle shipping address change requests by looking up the order, checking that it hasn't been dispatched, updating the delivery address if that's still possible, and telling the customer the outcome.”
  • CSAT survey. “After a conversation reaches a natural close, send a CSAT survey link by email that thanks the customer and invites them to rate the interaction.”

Improve and diagnose

  • “Review my setup and suggest specific improvements to its instructions, triggers, actions, and tests.”
  • “Help me diagnose why a ticket wasn't handled as expected. Ask me for the ticket number.”
  • “Why did the AI hand this ticket to a person?”

Run your workspace

  • “Help me customize my support dashboard around the metrics and queues my team needs.”
  • “Migrate our agents to Playbooks.”
  • “Help me understand how to set up this feature.”

In Slack

  • “How many tickets are open right now, and what are customers asking about most?”
  • “Send me a short report of yesterday's support every morning.”