Sequential questioning
Sequential questioning means asking one question at a time in a logical order, where each question builds on the previous answer. This helps your AI Agent guide users smoothly through multi-step tasks like returns, troubleshooting, or collecting order details, without overwhelming them. By breaking conversations into clear steps, the Agent stays focused, avoids confusion, ensures complete answers, and is better prepared to trigger actions or hand off to a human when needed.How to write sequential instructions
Use a clear list format in your instruction, so your AI Agent knows the correct order.Recommended instruction (Example: Return request)
When a user wants to return an item, ask for the following details one at a time:After all details are collected, add a closing sentence to manage expectations. E.g.:
- Order number
- Item to return
- Reason for the return
- Email address used for the order
- Zip code and house number of the delivery address
When you’ve received all data, let the user know we’ll send return instructions per email.You can also trigger an Action or Human handover once the data is complete. Learn how to do this here: Using Actions Human handover
Example of sequential question flow
Here’s how a return conversation could unfold:User: I’d like to return something. Agent: Sure — what’s your order number? User: 123456 Agent: Thanks — I’ve got order number 123456. Which item would you like to return? User: The orange Agent: Thanks — what’s the reason for the return? User: I prefer watermelon Agent: Got it — thanks. What email address did you use to place the order? User: [email protected] Agent: Thanks — I’ve got [email protected]. What zip code and house number was the order delivered to? User: 3511 AB, 12 Agent: Thanks — I’ve got the zip code and house number (3511 AB, 12). We’ll process your return for order 123456 (item: The orange, reason: I prefer watermelon) and send the return instructions to [email protected] shortly.