Practical AI for People Who Get Their Hands Dirty

Matt Shepard is a busy general contractor and real estate investor. He got his time and his sanity back by building his own AI virtual assistant. He had no tech background and taught himself, starting with systems for his own companies. Now he builds AI automations that run real businesses. In this conversation he explains how he decides what to hand off to AI and how much access to give it to things like your calendar and inbox. He also explains why documented SOPs and owner know-how make adopting AI so much easier. We also cover how he tests his systems, what AI means for entry-level jobs, the tools he prefers, when to reuse a proven workflow versus build something custom, and some of his most unusual projects. Matt closes with a message for business owners who are still unsure about AI. Visit Matt's website for more on how he can help you implement AI: https://www.shepardconsulting.ai/ Connect with Matt on LinkedIn:   / mattshepardaiguy   Visit Kuder Consulting Group website to learn how we help family business owners balance family and business: https://kuderconsultinggroup.com/ 00:00 What was it like to work with your great-grandfather on the job site? 04:16 What was the first AI system you built, and how did you approach teaching yourself to build it? 08:11 Was there any special learning curve in connecting AI to tools like your calendar and does that present additional risk? 14:03 Would it be simpler to implement AI systems if a business already has documented processes and SOPs in place? 15:33 Do you use AI to help you build the tool that the AI is going to use? 18:55 If AI takes over entry-level white-collar jobs, who will gain the experience needed to move up and eventually fill senior roles? 21:01 Is it okay to ramble when talking to an AI, and can giving more spoken context lead to better answers? 29:16 When you look at yourself and your business, how do you see yourself as different from the typical AI implementation consultant? 33:59 In your world, are you building from scratch for each client every time, or is there a combination of reusable parts and connectors? 36:01 How do you help potential clients bridge the gap between casually using AI tools like ChatGPT and actually figuring out what useful automations or applications they can build? 43:02 What's the most surprising use case you've built so far? 50:51 How can listeners reach you if they want help, and what would you say to a business owner who’s on the fence about bringing AI into their business for automation and admin?