After getting a yes from the client, what is the procedure, and what documents should I be ready with to share with the client?

When I started the AI Automation Society, I had no idea it would grow into this. Under two years later and we just crossed 400k. The largest AI automation community in the world. Yes, the space is exploding. But this community grew because you guys keep showing up, asking questions, dropping answers, sharing builds, and helping the person one step behind you. Huge thank you to the team that keeps this thing running, and to every single one of you who's posted, commented, or just been here. I feel lucky to get to do this. So excited to share with you guys what we've been quietly cooking up over here at AIS...👀 - Nate

But I pick mine based on feel. That probably sounds backwards because we're trained to trust the numbers. This model scores 90%, that one scores 80%, so the first one must be better... right? Well a story broke this week about SWE-Bench. It's the test that checks whether an AI can fix real problems in real software, like a human programmer would. It's the score a lot of technical people have leaned on for over a year. Turns out the models were cheating. The test projects already had the correct answers sitting inside them. So instead of solving the problem, a model could peek at the solution and hand it back. Like taking an exam with the answer key taped inside the textbook. On the SWE-Bench, GPT-5.5 scored 58.6% and Gemini 3.5 Flash scored 55.1%. Only 3.5 points apart? If you've ever used those two models, you know the math isn't "math-ing" there. Then a new test showed up called DeepSWE. Same idea but they pulled the answers out, so the model has to actually figure it out. On DeepSWE, GPT-5.5 scored 70%. Gemini 3.5 Flash scored 28%. The two "tied" models weren't close at all. And that gap lines up with how different these tools actually feel to use. None of this makes benchmarks useless. They're fun to look at and they give you a rough starting point. But remember who makes most of them. A big score is a marketing asset. It's the number on the launch tweet. The keynote slide. The headline. So always take them with a grain of salt. What I actually do is I bounce between Opus and GPT all day. Not because one won a benchmark, but because I've built a feel for which one handles which kind of task. For serious work right now, those two are the only horses I really trust in this race. Building that feel isn't exciting. You take one task you actually need done, run it through different models/harnesses, and notice which one you trust with the result. Do that enough times and you stop reaching for the leaderboard. → A model that's perfect for someone else can be the wrong pick for you.

I’ve been watching our 3,700 students in AIS+, and I noticed the people making the most money are all doing this: https://app.aiautomationsociety.ai/10-hours The people charging $5,000, $10,000, even $50,000 per engagement weren't better builders. Before they ever opened n8n/Claude Code, they did one thing differently: → They found the automations worth building first. It’s like a mini audit. Just by asking a few questions and mapping out the opportunities, they were able to get clients excited and also choose the right projects to work on. And the best part is you can practice by running this same system on YOURSELF. I call it 10 Hours to 10 Seconds, because doing this can easily save you or your clients 10 hours a week by automating the right things. Get all the details here: https://app.aiautomationsociety.ai/10-hours Talk soon, Nate PS: If you’re in AIS+, this has already been updated and provided to you at no cost. You can find it in the classroom

Can you guess how many prompts it took me to generated this UGC ad? 5 clips. 8 seconds. Zero edits That is all it took to generate this UGC using Google's new Gemini Omni Flash I know what you are thinking. You guessed 5 to 10 prompts minimum It was one. One detailed, well structured prompt I used a specialised skill built specifically for Omni UGC inside Claude to plan and write the prompt. Instead of guessing what to say, the skill handles the structure, the shot breakdown, the tone, and the pacing Once the prompt was ready, I handed it over to the Flow agent It spun up 5 generations simultaneously What came back were 5 raw clips I clipped them together. No cuts. No colour grading. No edits of any kind What you see is straight out of the model