An AI chat agent answers customer questions and hands off anything it cannot solve to your support team in Slack. It asks for an email first, so your team can follow up. Great for support teams that want fast handoff from bot to human with clear context.

Incoming chat messages start the agent. It runs on an OpenAI chat model with short term memory to keep the conversation on track. When the agent is unsure, it uses a custom tool that calls a sub workflow. That sub workflow checks if the user included an email. If yes, it posts the full message to a Slack channel and confirms that a human was notified. If not, it prompts the user to share an email so your team can reach them. This flow reduces dead ends, keeps the chat friendly, and moves the right items to people.

You need OpenAI and Slack accounts and valid credentials in n8n. Set the Slack channel, link the tool to the sub workflow, and map the chat input. Expect faster answers, fewer lost chats, and better follow up. Use it for pre sales questions, technical help, or after hours coverage where quick human escalation matters.

• Chat trigger listens for new messages and starts the AI conversation instantly.
• AI Agent powered by an OpenAI chat model responds to users with short term memory for context.
• Simple Memory window keeps recent messages so replies stay relevant.
• Custom tool Not sure calls a sub workflow when the agent lacks confidence.
• Execute Workflow Trigger passes the chat message into the sub workflow for checks.
• If node checks whether an email was provided and routes to Slack or an email prompt.
• Slack node posts the user message to a support channel for a human to help, then confirms back to the user.

• Reduce triage time from 30 minutes to 2 minutes by routing unclear questions straight to Slack
• Automate 80% of escalation steps including email capture and team alerts
• Capture contact details on 100% of escalations before a human steps in
• Handle up to 5 times more inquiries with the same team by removing manual handoffs
• Connect AI chat and Slack without copy and paste or context loss

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with OpenAI and Slack. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, double click the OpenAI Chat Model node, choose Create new credential, name it clearly, and paste your OpenAI API key from the OpenAI dashboard API keys page.
4. Double click the Slack node, choose Create new credential, and follow the on screen steps to authorize your Slack workspace. Make sure the bot has permission to post to channels and is added to your target channel.
5. In the Slack node, set the Channel field to the channel ID or name where your team wants alerts.
6. Open the AI Agent and confirm it is connected to the OpenAI Chat Model and the Simple Memory nodes.
7. Open the Not sure tool and confirm it is linked to the sub workflow. In the sub workflow, check that the Execute Workflow Trigger input name matches chatInput.
8. Open the If node in the sub workflow and confirm the condition that checks for an email value. Adjust the field path to match your chat input if needed.
9. Verify the Code node that prompts for an email sends a clear request to the user, and the confirm node replies when Slack has been notified.
10. Enable the chat trigger. Start a test chat and send a message the model cannot answer. For example, ask for details it was not trained on.
11. When asked, provide an email address. Check that Slack receives the message with the original user text and that the chat confirms a human was notified.
12. If Slack messages do not appear, confirm the Slack credential is authorized, the bot is in the channel, and the channel name or ID is correct.
13. If the sub workflow does not run, check the tool linkage, the Execute Workflow Trigger inputs, and the If node field mapping.