Hey everyone! 👋 I've been trying to get my head around Relevance AI for a few weeks now, mostly to automate some customer follow-ups. I hit a wall pretty fast because my basic plan only allows for a few concurrent tasks in the agent queue, and things were getting stuck.
I needed a way to line up tasks without paying for a higher tier just yet. Then I had this super simple (maybe obvious?) idea: using a public Google Sheet as an external task list that my agent can read from.
Here's the basic flow I set up:
1. I have a Google Sheet where I (or my other tools) append new tasks. Each row has details like a customer email and what follow-up they need.
2. I published the sheet to the web (as a CSV).
3. In Relevance AI, I made a simple agent that, on a schedule, fetches the CSV from that public link.
4. It processes the first few rows, runs the tasks (like drafting an email), and then I have a step that *deletes* the processed rows from the sheet using Google Apps Script (this part was a bit tricky to set up).
It's basically a DIY queue! It's not perfect and feels a bit clunky, but it lets me batch up tasks overnight without hitting concurrency limits. I was wondering if anyone else has tried something like this? Or is there a much smarter way to handle job queues on the starter plan that I'm totally missing?
I'm still new to all this, so any advice would be awesome. The hack works, but I'm sure there are pitfalls I haven't thought of yet.