Quick one on Grok Bot since a few of you asked after the access change last week. The mistake almost everyone makes in week one is building a row of specialists and then having to remember which one owns which job. Ten bots, ten chat windows, and you end up opening none of them. What works better is one front door. You build a single bot, give it a human name, and it's the only one you ever message. Its whole instruction is four lines: Before doing anything, check whether another bot owns this job. If one does, route it there. Only do it yourself if nothing on the bench fits. Bring the finished work back into this thread. Behind it you keep a small bench. Research, content, visuals, planning. Four is plenty to start with. The routing only works if each specialist describes itself in one sharp line, so keep those descriptions narrow. A few things worth knowing before you build it: Interview yourself first. Hand the bot a file that tells it to grill you about your business and what you're actually trying to do this quarter, and tell it not to stop at your first answer. An interview pulls out things you'd never think to write into a system prompt. Have it save that as its own reusable skill so later you can just say "grill me about this plan". Let the chief hire its own team. It can already see which bots exist, so show it your goals and ask what roles are missing. And when your bottleneck moves, tell it what you're stuck on and let it restructure the bench rather than bolting on another generic helper. Collecting bots is the real failure mode. Ten you never message is worse than three you do. Be deliberate about which shelf a fact goes on. Shared knowledge is the stuff every bot needs: what your business is, who's on the team, your channels and offers, which tools you use. Private to one bot is working notes and half finished reasoning. As the team grows, misplaced context is what starts making agents feel dumb. Make them log the work. Have the chief open a task before it delegates, then have the specialist write its progress and a link to the output into it. Otherwise results disappear into a scroll of chat you never reopen.

Welcome to the AI Untapped Workshop. This is the room behind the newsletter and the YouTube channel. The whole point: cut through the AI noise, faster, together. Here's what lives here: → One new tool + prompt every Sunday. Posted in the feed. Copy, paste, run it. → The Field Notes: five short PDFs covering local AI setup, prompt patterns, the skip list, the maker stack, and the Sunday routine. Find them in the Classroom tab (top nav) or scroll the feed today, where each one is posted separately so you can ask questions in the comments. → An open thread where you drop the AI tools you're testing and what's actually working. → Behind-the-scenes context I don't put on YouTube. What I'm asking from you: → Post what you're testing. Wins and failures both. The room learns faster when we're honest. → No pitching. No "I built an agency, DM me." This is a workshop, not a marketplace. → If a prompt works, drop the result in the comments. If it breaks, drop the error. I'm here daily. Drop a hello below and tell me one thing you're trying to figure out with AI right now. I read every reply. — Dusty