Most AI Advice for Agencies Is Noise. Here's What Isn't.

Alex and Chris talk through how agency owners can stop feeling overwhelmed by every new AI tool and tactic, and focus on the small handful of things that actually move the needle in 2026. They get into why MCP is the one piece of AI infrastructure worth adopting now, why llms.txt and an "info for AIs" page can change how AI tools recommend you, how Claude Skills on GitHub are quietly becoming one of the best lead magnets for agencies, and where AI still can't be trusted (hint: anything client-facing). They also share how they built a free agency analytics tool inside Wayfront in days, replacing what used to take months.

00:00 Are agencies falling behind on AI?

02:17 The 80% of AI advice that's already noise

04:04 Why MCP is the only AI tech worth adopting right now

08:29 The agency pricing problem: seats vs value

12:43 Solopreneurs are the new mini-agencies

15:12 AI automations vs traditional automations (and why AI ignores you)

16:02 Wayfront's Kanban + natural-language rule builder

23:47 llms.txt and the "info for AIs" page that changes how AI recommends you

26:53 The Claude Skills lead magnet trend (and how agencies are using it)

33:02 Buddy is smart but slow: the speed problem with AI agents

38:28 We built an agency analytics tool in days, not months

40:49 The "5 things cooking at once" workflow shift

43:41 Working from a phone with a keyboard

Show notes