AI is not “coming” anymore, it’s already here, and the real advantage is shrinking fast. Robbie and George talk about why 2026 feels like a tipping point for AI automation, AI agents, and the pace of change across business and society. The conversation keeps returning to a simple idea: the early-mover window closes the same way it did with social media. If you wait until the path is paved, you’re no longer early, you’re competing with everyone. That’s why “playing with ChatGPT” isn’t the point. The point is building repeatable leverage in your operations before competitors do, whether you run a construction business, a trade company, or any service-based small business.
George explains what pushes him from curiosity to urgency: admin load, more responsibilities, and the realization that high-value leaders get dragged down by low-value tasks. They break down what an AI agent actually changes compared to using a chatbot as a search engine. An agent can be connected to tools like email, calendars, documents, CRMs, and task apps, then execute workflows, not just answer questions. The example is painfully relatable: searching for a vehicle registration or a loan payout figure buried in old emails. With the right setup, Claude or ChatGPT can retrieve information in seconds, and the “brain” of the agent grows as you add permissions, integrations, and clear instructions. The practical takeaway is that AI productivity is less about magic and more about plumbing.
A major theme is workforce disruption and the new definition of “employable.” Robbie argues you won’t get replaced by AI itself, but by someone who uses AI to produce three times the output. That shows up in hiring too: “AI savvy” becomes non-negotiable, meaning willing to use AI tools, learn fast, and apply them to real work. They also point to clear market signals like layoffs at large companies and internal AI questionnaires that test whether employees can prompt effectively. For business owners, the lesson is direct: you don’t need an abstract AI strategy, you need an AI roadmap that maps repetitive tasks to automations. Reporting, invoicing, scheduling, content workflows, and email handling are all high-ROI starting points because they happen every week.
They also tackle the objections that slow adoption: privacy fears, data access, and the worry that an agent will do the wrong thing. The hosts call out the inconsistency of trusting big platforms with your inbox while refusing an AI assistant the same access. They discuss safer adoption patterns like giving an agent limited permissions and a spending cap, similar to delegating to a human assistant. On the bigger “doom and gloom” question, they land on a grounded view: if something goes wrong, it’s likely human-driven misuse, not a movie-style AI rebellion. Either way, refusing to learn AI doesn’t protect you, it only reduces your options. The best move is to document workflows, get your systems out of your head, start with one annoying task, and build momentum from there.