Alright… most marketing teams right now are using AI. The real question is if you're using it in an order that goes anywhere. Because that's where it falls apart. You read about one tool for ads, another for email, a third for landing pages — and you end up with ten disconnected tricks and no system. Nothing stacks. The Playbook fixes that by laying marketing out as ten jobs AI actually does, in the order they feed each other. Here's the shape of it: Foundation first — know the buyer, build the content engine. Everything downstream feeds off these two. Skip them and the rest is guesswork. Then distribution — LinkedIn, email, ads. Getting the thing in front of people. Then conversion — landing copy built around the objection, sales conversations, getting named when someone asks an AI "what's the best X". Then the loop — one number per job, reviewed monthly, and last month's customer words rewriting next month's persona. The order matters more than any single tactic. Buyer, then content, then one channel, then conversion, then measure. Never start at ads. Never start at tools. That's the mistake that keeps people busy and broke. The full playbook, with all ten jobs laid out properly, lives here: https://www.skool.com/ai-for-beginners/classroom/4065432a?md=27b1e38e6cd54f1ba4ff417a69583737 And if you want the whole system on one page — the ten jobs, the order of operations, the four rules that make any of it work — start here: https://www.skool.com/ai-for-beginners/classroom/25737100?md=808cd999e665488c9ac462e6c044dd0b

Alright… this one's for anyone who's used AI a few times, felt the spark, and then kind of stalled. No judgment, that's most of us. The wall is almost never the tools. It's not knowing what to do next. So you open twenty tabs, everyone's shouting about something different, and nothing actually gets finished. The twelve-month roadmap fixes that by turning it into five phases. One step after the other, not a map you stare at. Here's the shape of it: Months 1-2 — work out where to point it. Score your tasks, draw the simple impact map, and hand ONE task over properly. Success here is just a couple hours a week back. That's it. Months 3-4 — become a power user of one thing. One assistant, your own briefs saved, one piece of content turned into five. You stop re-explaining your business every single time. Months 5-6 — get your numbers out of hiding. Pick three or four that matter, into one sheet, updated without you. Months 7-9 — fix the communication loop. Meetings and enquiries handled so nothing falls through the cracks. Months 10-12 — connect it up. The pieces start talking to each other, but only for the stuff you've already proven by hand. That last bit is the part people skip and regret. You don't automate a mess. You prove it manually first, then wire it together. The whole roadmap, with the phases laid out properly, lives here: https://www.skool.com/ai-for-beginners/classroom/20abb1f4?md=96c9f3db0c0f4bb697507ce16f8e5745 And if you're stuck on phase one — which task to actually hand over first — this is the lesson that walks you through scoring them so you pick a real win and not a three-month distraction: https://www.skool.com/ai-for-beginners/classroom/20abb1f4?md=b51537c33505403ea4588e5c59bfe4ce One thing to know: a couple of the older videos inside these lessons name tools and plans that have moved since. Follow the method — that's the part that lasts — and check the current screen if it doesn't match.

Alright… new one's live, and it's about Claude Cowork — the exact prompts I actually use, not a feature tour. Here's the thing. Most people open Claude, ask it to write something, get a wall of text, then have to go build with that text themselves. Cowork flips that. You point it at real work and it does the work with you, in the same place, instead of handing you a draft to babysit. So the video is me walking through the prompts I lean on. The wording matters more than people think. A vague ask gets a vague result. A specific one — real names, real context, the actual thing you want out the other end — gets something you can use. Watch it here: https://www.youtube.com/watch?v=tEkGwHeDPPo Who it's for: anyone already using ChatGPT or Claude a bit who wants to stop copy-pasting text around and actually get finished work out. If you're brand new to AI, still worth a watch — just know it clicks faster once you've spent a little time in a chat tool first. If you want to go deeper after, there's a lesson in the classroom on asking AI for the finished thing instead of the text — same idea, walked end to end: https://www.skool.com/ai-for-beginners/classroom/283e1613?md=78e7366317f340788d34147fd75679e7 One thing to know: model names and menus move fast in this space. Follow the method — that's the part that lasts — and check the current screen if it doesn't quite match Update: the full step-by-step for this video is now in the classroom — every prompt from it, ready to copy: https://www.skool.com/ai-for-beginners/classroom/c93a5c55?md=ec24e14e89ff47ec813230fd42a202b7

Alright… new one just went live, and it's about making creator-style video ads without hiring a creator. Here's the thing about UGC ads. The casual, filmed-in-my-kitchen kind of video outperforms polished production on every short-form platform. Everyone knows that by now. The problem was never the idea — it was making enough of them. A good UGC creator runs a few hundred per video and takes a week of back-and-forth. And you don't need one video, you need ten. Different hooks, different angles, backups for when one stops working. That's the part this new stuff cracks open. You take a product photo off your phone and generate a batch of realistic ads around it — fast enough to actually test which one lands. Watch it here: https://www.youtube.com/watch?v=kwysV2smgfY Who it's for: anyone selling a physical product who wants to run ads and doesn't have a creative team. Works best with physical products — if you sell something digital, use a screenshot or a mockup and expect to test a couple of versions. If you want to go deeper on the workflow after watching, there are two lessons in the classroom that walk it end to end. Making a batch, different hooks and avatars, start to finish: https://www.skool.com/ai-for-beginners/classroom/39633a7c?md=48efc91559dd40aaa08aa6a609444771 And the TikTok reference-video method — point at a video that already went viral and rebuild that structure around your product: https://www.skool.com/ai-for-beginners/classroom/39633a7c?md=ee37138780064f2a845c446ce28857aa One thing to know: the videos move fast and model names change every few weeks. Follow the method — that's the part that lasts — and check the current screen if it doesn't quite match Update: the full step-by-step for this video is now in the classroom — every prompt from it, ready to copy:

Alright… this one's for anyone who's typed a lovely long sentence into Midjourney and got back something that looks nothing like what was in their head. Here's the thing most people get wrong. Midjourney is not ChatGPT with pictures. That's the most expensive misunderstanding out there. ChatGPT reads your sentence and tries to understand what you MEANT. Midjourney doesn't. It weights the words you gave it and builds from those. It understands nothing — it just reacts to what's there. So the fix isn't writing more. It's writing the RIGHT things, in the right order, and letting the parameters do the heavy lifting. 🔥 What actually moves the needle: ➡️ Say the subject first — the thing you want, plainly, before anything else ➡️ Then the setting, the light, the mood — one clear detail each, not a wall ➡️ Then the parameters — that's where 90% of the work happens A couple that earn their place every time: --ar sets the shape. 16:9 for a wide thumbnail, 3:2 for a product shot, 1:1 for a square post. --style raw pulls it back toward realistic and away from the glossy default look. Great for product and brand work. --s (stylize) is how much creative liberty it takes. Low for clean and literal, high for arty. One more that saves you an hour: never ask Midjourney to put words in the image. It mangles text. Generate it clean, then add your headline in Canva after. The method that makes it click — stop rerolling fifteen times hoping. Write one line, look at what came back, then nudge one thing. Change the light, or the angle, or the stylize number. Refine, don't restart. The full lesson walks through the exact structure with real prompts you can copy — product shots, brand illustrations, the lot: 👉 https://www.skool.com/ai-for-beginners/classroom/ed9ee3da?md=00db8c220c22460694397166e8e17af9 One note as always. Midjourney moves fast — the interface and parameters shift, and some of the video looks a little different now. Follow the method, that's the part that lasts, and check the current screen if what you see doesn't match.