i need help as my whatsapp is executing perfectly but the issue is that when the AI responds to a message it sends a webhook link but does not talk any further. what mistake am i making ?
The issue is that your AI node is sending a message back but the WhatsApp trigger isn’t set up to handle the conversation state correctly. When your webhook fires and the AI replies, that reply goes back out – but the trigger doesn’t automatically “listen” for the next incoming message from that same user unless you have a loop or a persistent session structure in place.
A few things to check:
1. Is your webhook set to respond vs. continue?
If you’re using the “Respond to Webhook” node to send the AI reply, n8n considers that workflow execution done. Any follow-up WhatsApp message from the user will fire a fresh execution – and if your workflow doesn’t handle session continuity, it’ll just reply once and stop.
2. Are you storing conversation history?
For a back-and-forth conversation, you need to save the message history somewhere (a simple Google Sheet, a database, or even n8n’s built-in memory nodes if you’re on a recent version). On each new incoming message, load the previous turns and pass them to the AI. Without this, the AI treats every message as a brand-new conversation.
3. Check your WhatsApp node configuration
If you’re using the WhatsApp Business Cloud node, make sure you’re only triggering on “message” events (not status updates like “delivered” or “read”) – those status webhooks can cause extra executions that look like the flow is broken.
Quick fix to test: add a simple IF node right after the trigger that checks if $json.entry[0].changes[0].value.messages actually exists and is non-empty. Status callbacks won’t have that field and will be getting caught by your AI node unnecessarily.
What does your trigger node’s incoming data look like? If you share the structure of the webhook payload you’re receiving, it’s easy to pinpoint exactly where it’s cutting off.
Hi @Azwer_Afzal this can happen if your flow is not sending the AI agent’s output back to the user assuming that you have a pattern something like this:
WhatsApp Trigger → AI Agent → WhatsApp Send Message node
Also make sure to have an IF node so that the messages that actually contain some text will be passed to the AI agent.
Also make sure you are using this {{ $json.output }} at the end of the flow which sends the user message curated by AI also consider refreshing those credentials if that helps.
1 Like
same issue is coming up again
@Azwer_Afzal can you share your workflow file?