Most people want to skip the setup and jump to the fun build. I get it. But the first episode of this series builds nothing at all, and I did that on purpose. You put one tool on your computer, Claude Desktop, and you get Claude Code ready inside it. That's it. Every episode after this one runs on that same tool, so an hour spent getting it right today saves you a pile of confusion later. Here's the line from the clip that sums it up: No building in this episode, just set up, done once. So before we go further: are you set up already, or is this the step you keep putting off? Tell me where you're stuck and I'll point you at the fix.

Reading about an AI receptionist and watching one book an appointment are two different things. So here's the second one. We loaded our academy details into a voice agent and let it take a booking call. It offered the kids jiujitsu program, asked the ages, found open days, and locked Evan into a Saturday trial. Nobody on our side picked up. The part that surprised me wasn't the words. It was the pacing. It waited, it confirmed, it moved on like a good front desk person would. Where would a call like this fall apart for your business?

Someone's about to sell you an "AI agent" for a few grand and hand you a chatbot. Same word, wildly different thing, and the gap is your money. The good news is there's one question that exposes it on the spot, and most sellers fold the second you ask it. This week's episode hands you the whole BS detector for buying AI. 🎯 The one thing Make them prove it. Whatever they're selling, ask them to show you a real system doing the actual work, live, right now. Their own, or a client's. The real ones light up and show you. The pretenders suddenly only have slides and a follow-up call. Risk read: 🟡 use with caution. The tech itself is fine. The risk is the purchase, paying agent money for chatbot work, or handing an "autonomous" tool the keys on day one. If you do one thing this week: 1. Before your next AI pitch, decide what you actually need: does it just need to TALK, or does it need to DO real work? That one answer sorts chatbot vs assistant vs agent, and roughly what it's worth. 2. Ask the seller to show you something real and working, live. Watch what they do when you ask. 3. Start small. Scope one slice of your business, prove it works, then expand. If the only option is all-in up front, that's a flag. What's the AI pitch line that's made you most suspicious as an owner? Drop it below and we'll break it down 👇 🗳️ Quick poll: Have you ever paid for "AI" and felt like you got less than you were promised?

Most people ask if AI voice agents are legal. Wrong question. Ask which kind you're running. Here's the split that sets your risk. Answering inbound calls, where the customer chose to call you, is the lower risk lane. That's the receptionist case, and it's where most small businesses should start. Making outbound calls is a different world. That's where the law gets heavy and where people get themselves in trouble. Same tech. Two very different rulebooks depending on who dialed. Are you thinking inbound, outbound, or both? Want to know where people are actually starting.

Here's the moment this clicked for me. It's 9pm. My team's home. A parent calls about a trial class. Old way, that's a voicemail nobody hears until morning, and half of those people already booked somewhere else by then. New way, a voice agent answers, gives a real answer, and books the appointment. I'm at dinner with my family and I never touch the phone. After hours is where most owners feel this before anything else. The lost calls were never dramatic. They just quietly left. How many after hours calls do you think you're losing right now?