Chatfuel

Why Fuely Replies to Each Message Separately

Fuely AI replies to every incoming message individually by default.

Last updated on

Fuely AI replies to every incoming message individually by default. If your customers tend to send several short messages in a row, you can turn on message delay so that Fuely waits and responds to the full thought at once.

Even with message delay enabled, you may still see separate replies. The most likely reason is the time gap between the customer's messages.

How message delay works

When message delay is off, Fuely AI generates a reply to every message as soon as it arrives — each message gets its own response.

When message delay is on, Fuely AI waits 8 to 12 seconds after receiving a message before generating a reply. During this window it collects all incoming messages and responds to them as a single batch.

If the customer sends their next message within that 8–12 second window, Fuely bundles it with the previous ones and replies once.

If the customer sends their next message after the window has closed, Fuely has already started (or finished) generating a reply to the earlier messages. The new message is treated as a fresh input and gets its own response.

Why it can look like the delay is not working

Customers sometimes pause for 15–20 seconds (or longer) between messages — for example, while thinking or switching apps. That pause exceeds the delay window, so Fuely replies to whatever it has received so far. When the customer sends the next message, Fuely replies again. From the customer's perspective it looks like the bot is "not waiting," but it did wait the full 8–12 seconds each time.

What you can do

  • Let customers know the bot is listening. The typing animation shows automatically during the delay, signaling that a reply is coming. This encourages customers to send all their thoughts before the window closes.
  • Check the timestamps. In the chat log, compare the time of the customer's messages. If there is more than ~12 seconds between them, the separate replies are expected behavior — not a bug.

On this page