So my AI Agent calls a tool that sends info to a webhook module in Make.com and I set everything up perfectly. The only issue is sometimes if I haven’t spoken to the AI Agent in a while and instructed it to call the tool it won’t work the first time. For example, I hadn’t talked to the AI Agent in about 24 hours and I asked it to call the tool and it didn’t work (no data was sent to Make.com). Then I tried again right after and it worked and it worked subsequently for the rest of the night. I did not change anything and this has happened multiple times. Do you know how to solve this or if this is a typical problem people run into? Thanks for your help!

The webhooks may fail after inactivity due to webhook timeout or cold starts, ensure the webhook scenario in Make is always active, send periodic test triggers, and consider adding retries or error handling in your tool logic to maintain reliability..

Thanks for the suggestions and yea I made sure the webhook in make is always active. Is there a way to bypass this problem without needing to send a daily test trigger to keep it warm? Also I was trying to add retry logic to the tool but it looks like you can only add retry logic to JavaScripts not JSON. So my next thought was to add retry logic to the AI Agent’s actual prompt. Would that fix it and if so how should I structure it?

Yeah the retry logic only works in JavaScript, but you can instead guide the AI agent via its prompt

Can I share my current prompt with you and ask for your advice on how to structure the retry logic within the prompt? Also if there’s another better solution that you know of can you please advise me. Thanks for your help!

You can share your prompt