I'm currently doing cold outreach to business owners for my AI automation consulting business. My goal isn't to sell AI solutions immediately. I want to understand their operations, identify real bottlenecks, and see whether there's an opportunity to help. I'm comfortable making cold calls and reaching decision-makers, but I'm trying to improve the transition into discovery. My question is: When you reach the business owner, do you start a short discovery conversation right there on the cold call, or do you briefly explain what you do and schedule a separate 15-minute discovery call? And how do you explain the value of that discovery call without turning it into a sales pitch? I'd appreciate hearing what actually works for you.

“When starting an AI automation agency, what type of automation should I build and sell first: a traditional workflow automation, or an AI-powered automation that uses LLMs/RAG?”

Just booked my second free business audit for October 13th at 1 PM! This one came from a cold call. ✅ Made the cold call ✅ Got the prospect interested ✅ Sent over a questionnaire to understand their business and pain points ✅ Scheduled the Google Meet Now the goal is to prepare a valuable, personalized audit, identify opportunities where AI and automation can actually help, and hopefully turn it into a great case study. Still early in the journey, but it's good to see that taking action is starting to create opportunities. A little nervous, but excited. Hopefully it goes well! One call at a time.

Just closed my first full AI voice agent project for a restaurant. The client will pay €1,000 now and another €1,000 in 3 months, so €2,000 total for the setup. The project was for Sushi Maison Boltiere, an Italian restaurant. I built an AI phone receptionist that can: - Answer incoming calls naturally - Check table availability - Book reservations - Modify or cancel existing bookings - Send WhatsApp confirmations automatically - Update a CRM with the reservation data and changes The system is built mainly with Retell AI + n8n, connected to the restaurant's booking platform, with Telnyx handling the WhatsApp side. The biggest part of the work wasn't just building the workflows — it was testing everything properly. I ran hundreds of test calls and worked through edge cases like changing dates/times, unclear customer requests, duplicate checks, cancellations, large groups, notes, and making sure the AI didn't repeat or ask for information it already had. It took much longer than I originally expected, but the project is now finished and ready for the client. Definitely learned a lot from actually building something that has to work with real customers, not just in a demo. €2K project closed. Now onto the next one.

Two updates from the restaurant niche I've been working: Win #1 — Sold an AI-built menu. Closed and delivered a digital multilingual menu for a restaurant client, built with AI. First paid deliverable in the pocket. Win #2 — Same client, bigger scope. Because they were happy, they came back and asked me to build them a reservation management system + a voice agent to handle bookings. Ran my first audit on their current setup, and I'm sending over the report today. Biggest lesson so far: land the small win, deliver clean, and the upsell comes to you. One good menu turned into a full system + voice agent conversation.