In this video, I walk you through how to make your n8n RAG agents instantly smarter using re-rankers and metadata. If you've ever had your AI agent return the wrong information or pull irrelevant chunks from your vector database, this is the fix. Re-rankers help prioritize the most relevant chunks after your initial search, making the results dramatically more accurate. I also show you how to use metadata both when vectorizing and when retrieving, so your agents can filter and find exactly what they need with precision. These techniques are simple to implement but make a huge difference in how well your agents understand and respond. If you're building AI agents in n8n, this video will help you take them to the next level. Must be on version 1.98 or higher to access rerankers.

Just thought I’d make a quick post to give you some ideas on how AI agents can actually be useful right now—especially if you run a Skool community or do growth work. Been experimenting with n8n (pretty much same thing that I did last time with make.com) and built a little AI setup that turned out to be kinda fun and actually practical. I usually don’t trust AI agents for stuff like email replies or booking tasks—they fail often. But if you use them as smart assistants that pull context and help you understand things? That’s where it gets interesting. Here's what I built: - A basic HTTP Request node that pulls the latest Skool data (posts + community stats) - Data goes into a Google Sheet to track history and compare trends - Hooked up an AI agent that answers questions like: - The agent grabs the live data, checks trends, and replies in plain English (or in your language) - You can trigger it on-demand with the n8n chat trigger It's basically a Skool-aware GPT assistant you can talk to. Helps you get a good data snapshot and understanding your community's healt. Getting the signal from the noise. Hope this helps! 🥊

Been playing with an idea: AI agent turns community call transcripts into weekly Skool posts or newsletters. Built on Apify/Other relevant tool for calls + n8n AI Agent + Skool API. Worth pursuing?

Is this kind of stuff interesting to you? Let's chat here. Want better visibility into your AI workflows in n8n? In this video, I’ll show you how to track the actions your agent takes, log the outputs, monitor token usage, and calculate the cost of each run. This kind of observability is key when working with AI agents, especially to catch hallucinations and control costs, so you can continuously iterate on the system prompt. If you’re building agentic systems in n8n, this will help you stay in control. Google Sheet Template

For a deeper discussion on building AI systems that provide meaningful ROI, let's chat here. In this video, I’ll show you how I’ve been saving hours each week by using n8n and OpenAI’s image generation API to automate marketing visuals. It’s surprisingly easy to set up — and in this walkthrough, I break it down step by step so you can follow along, even if you’re new to n8n or automation in general. If you're doing a lot of manual design work or just curious how AI can help streamline your workflow, this is a great place to start.