I ran a small experiment today: used Claude Skills to spin up a few “digital employees” and hand off daily tasks to agents. The idea was inspired by Greg Isenberg’s video—I adapted it, tested it, and it actually works really well. What I built: an A/B idea agent, a marketing insight analyzer, and a tweet → newsletter converter. Setup was quick and the results were clear. If you want the high-level flow or my notes, I’m happy to share.

OK so this started off as a workflow, but after the conversation with Nick Sonnenberg and @Nate Herk today, I thought I would try something a bit different. Nick asked if Nate had taken a workflow he had created and tried to turn it into Code. So I thought I would try it! Here is the result: https://n8ndocumenter.replit.app/ Scenario: You have a workflow you have downloaded but you don't understand how it works, OR you have written a new workflow and you didn't have time to document it. This tool will allow you to upload the workflow file or JSON and it will write you documentation and tell you what the workflow is about and how it works. It is version 0.1 so lots more to come. I have released this as another Community tool, but since it is using an AI API, I may have to reconsider it if it gets too popular and costs me too much money!

Hello 👋 What are the best ways to train AI Agent in similar way like a Custom GPT or a Project? I have several Custom GPTs and Projects with extensive knowledge which includes documents, links and special instructions and these are doing well . Now I would like to build automation with Agents , I need those to be trained as well. All help much appreciated

Hi all, I'm interested in setting up a custom agent for a content marketing workflow. LinkedIn, Instagram, etc. Obviously, in order to get the best results I want to provide good direction both on 1) my own voice and tone and 2) examples of high-performing posts in each channel. For this use-case, would RAG or fine-tuning be a better approach for getting a content marketing agent up to speed?

Most people think they know how to prompt… Until they realise their “prompting” is really just hoping for good output. ❌ Vague commands. ❌ Random hallucinations. ❌ Wasting time rewriting the same thing 4 times. Sound familiar? We just launched Prompt Engineering Mastery: From Foundations to Future - a complete course on how to actually think and work with language models. It's 45 lessons, 7 quizzes, and 100% free. No paywall, no bs. Built for real-world use: 👉 Professionals 👉 Builders 👉 Teams 👉 Even total beginners It goes way beyond “how to write a good prompt”. You’ll learn how to structure, test, and scale prompts. How to work with agents. How to avoid garbage output. How to use prompts for research, decision-making, support, even product design. And yes - you can do it from your phone. This isn’t some fluff mini-course. It’s the most practical prompt engineering course we’ve ever seen - and you’ll walk away with a real skill. STRONG recommendation, even if you're already a veteran. Start here: norai.fi If you’re building with AI (or even just using ChatGPT for your job), this is the course you don’t want to miss. Level up. No excuses now. ps, took me days to do this. 🥵 Please leave a review afterwards.