219: "Why You Cannot Shortcut The Human Part Of AI Implementation" ft. Justin Coats

Justin Coats breaks down the idea of “super apps” for AI and what that really means for corporate adoption, change management, security, and tokenomics.

🧭 Conversation Highlights

• Justin defines super apps as AI workspaces that combine model access, agents, and computer or browser use in one interface.
• Erik pushes on the human side of adoption and asks whether waiting for the tech to mature actually reduces the real cost.
• They discuss how change management needs education, executive buy-in, and AI champion teams that represent more than just IT.
• Justin explains context windows (AI attention span) and then tokenomics: tokens measure workload, credits measure billing, and measurement is still messy.

💡 Key Takeaways

• Super apps are less about a new tool and more about consolidating how work happens, which forces a new way to interact with your systems.
• Waiting for “fully baked” AI does not remove the human adoption problem. Humans are slower, so competitors build advantage while you stand by.
• Good adoption looks like: leadership using it, education for everyone, and an AI champion group embedded across functions.
• Tokenomics and context windows are becoming operational concerns. If you cannot measure and manage usage, costs and performance turn into guesswork.

❓ Questions That Mattered

• Will adopting super apps get easier from a change management standpoint, or will it keep demanding constant iteration from teams?
• How should companies think about their tech stack being partially commoditized and what pushback comes from the original tools?
• If a super app depends on a small number of operators, what happens when key people leave and the wiring breaks?
• What does it mean when an AI says the chat is too long, and how should teams manage context and summarize threads?

🗣️ Notable Quotes

• “Humans need their hand held until they don't.”
• “Context window is your AI's attention span.”
• “Tokens measure the underlying AI model's workload. Credits are the product's own usage and billing currency.”

🔗 Links & Resources

• Listen To Other Episodes Co-Hosted With Justin
• Read the Article 'What a Forgotten 100-Year-Old Government Report Says About Who We Are'