Hi everyone, I’m Claire, a senior CS major in Korea with about a year left until graduation. Up until a year ago, I was on the path to becoming a Frontend Developer. But after seeing AI's capabilities, I’ve decided to pivot into AI Automation Freelancing😂 I’ve been diving into this world for about two weeks now, mostly inspired by Nate Herk’s content. My schedule is a bit lighter now that my full-time internship is over, so I’m ready to go all-in on learning and monetizing this. But honestly, I’ve hit a bit of a wall regarding my next step.. Here are the two things I’m really struggling with: 1. The "n8n vs. Agentic" Dilemma Looking at Nate’s videos, it seems like AI agentic workflows (like Claude Code) are the future. However, when I look at marketplaces like Fiverr, everything is still heavily centered around "n8n automation." In 2026, is mastering n8n still worth it? Or should I focus purely on identifying industry problems and solving them via Claude Code/custom workflows?? 2. The Delivery/Hand-off Process If I go the custom agentic route instead of using a low-code tool like n8n, how do I actually "deliver" the final product to a non-technical client? - Should I be building a custom UI wrapper or a dashboard where they can input their own API keys? - What’s the industry standard for handing off these kinds of custom AI solutions? I’ve been thinking this for a week and feel like I’m losing my sense of direction. I’m ready to put in the work—I just want to make sure I’m climbing the right mountain. Would love to hear any insights perspectives from this community. Thanks a lot in advance! 🫶

I automated my lead generation workflow. Here's what broke (and what I learned fixing it) 😅 The goal: Search Google Maps for businesses → Scrape websites for emails → Verify them → Save to Google Sheets. All automated. Sounds simple, right? It wasn't. Reality: Nothing worked the first time. I watched a YouTube video few months earlier and was trying to find a way to find emails of businesses for FREE. What broke (and what I learned): 🔴 Problem: I tried Apify but the limits were Hit in 2 days. Fix: I built an email pattern generator (info@, contact@, sales@) Lesson: Check API limits BEFORE you build 🔴 Problem: Getting emails like [email protected] (lol) My pattern generator got emails with www. Fix: One line code fixed "www." Lesson: Test with 1 item first, not 50 🔴 Problem: "Too Many Requests" errors everywhere Fix: Added a loop with 1-second delays between requests Lesson: Rate limiting is about workflow structure, not just code 🔴 Problem: All my data disappeared after API calls Fix: I used set data node to preserve data Lesson: Data doesn't magically stick around—you have to design for it 🔴 Problem: Paid API's everywhere 💀 The main problem with lead generations is to get verified emails and these are not FREE. Fix: Cut patterns from 15 to 6 per business (still works!) Lesson: Free tiers force you to optimize (not a bad thing) The real takeaway? Not every automation project is a portfolio piece. Sometimes the real value is in what broke and how you fixed it. After all the fixes: I can now generate 100-200 verified leads in 15 days, fully automated. But honestly? The debugging taught me more than the final result. Quick question: How you handle paid APIs costs