260: "Practical Use Cases for ChatGPT's Dots" ft. Justin Coats

Erik and Justin catch up on how voice-first AI agents like Dots (Q) are changing daily execution, then zoom out to the real adoption gap. They unpack why many MSPs and businesses are hesitant, what incentives are missing, and why the next wave is about layered context, specialization, and human guidance at scale.

🧭 Conversation Highlights

• Erik demonstrates “Q” as a voice agent: drafting emails, building decks, creating shared folders, and even researching and scaffolding a repeatable pre/post survey system.
• Justin explains how their agent setup uses specialized sub-agents, cloud virtual machines, and strict approval guardrails so the agent can work without going rogue.
• They debate the real risk conversation: AI is reducing button-clicking work, but raising the value of taste, quality, and domain judgment, often creating demand for experts and fractionals.
• Justin shares what he heard at an MSP/IT conference: most providers are not training teams on AI or supporting customers, mostly because the forcing functions are not yet landing.

💡 Key Takeaways

• Voice agents are most valuable when you already have structured communication in your history, because the agent learns your context and style from that access.
• Effective prompting often means going deep in context (layering information), not trying to “be clever” in the prompt.
• AI lowers the cost of high-quality execution for the long tail of businesses, which shifts work from “who can click” to “who knows what good looks like.”
• Adoption isn’t blocked by lack of interest. It is blocked by missing business incentives and lack of support infrastructure, especially among MSPs that lack training and a clear adoption roadmap.

❓ Questions That Mattered

• How much context can you reliably articulate to an agent, and how much does the agent need access to your real communication history to succeed?
• When businesses say they are “not doing anything with AI,” what are they actually missing that would create a real forcing function?
• Will AI replace professionals, or does it mostly reallocate work toward higher judgment and specialized expertise?
• How should teams standardize agent behavior so work stays secure, consistent, and scalable across staff and systems?

🗣️ Notable Quotes

• “Three weeks in AI time is like it’s been a year.”
• “Dots has access to my email… She sees every form of communication I have and hears it all.”
• “Nearly every single one of them said I’m not doing anything with AI. My team has not had training.”
• “The value of the human button clicking has gone down. What’s gone up is the value of quality, of style.”

🔗 Links & Resources

• Listen To Other Episodes Co-Hosted With Justin