His Own Ai Disagrees with Him 😳😳

Why It Matters

By embedding a confidential, always‑on AI coach, firms can multiply expertise, cut training costs, and accelerate decision‑making, giving them a competitive edge in knowledge‑intensive markets.

Key Takeaways

• Alex runs on Anthropic Claude, trained on firm’s proprietary data
• No internet access; answers solely from internal knowledge base
• Provides 24/7 confidential coaching, supplementing human coaches through AI
• Can disagree and synthesize, offering novel perspectives to users
• Scalable knowledge sharing promises future AI‑driven best‑practice repository

Summary

The video introduces Alex, an internal AI coach built on Anthropic’s Claude model and fed with millions of words of the firm’s proprietary content. Unlike public chatbots, Alex has no internet connectivity and draws answers exclusively from this curated knowledge base, ensuring confidentiality and relevance.

Alex offers 24/7 access to the firm’s best practices, acting as a supplement rather than a replacement for human coaches. It can challenge users, synthesize information faster than a person, and generate original insights while staying consistent with the firm’s thinking.

The presenter recounts a moment when Alex replied, “I respectfully disagree,” after being told it felt like talking to himself, highlighting the system’s ability to push back and provoke deeper analysis. Such interactions illustrate how the AI can surface unexpected yet aligned perspectives.

If scaled, Alex could democratize expertise across the organization, reduce reliance on individual coaches, and accelerate the dissemination of thousands of best‑practice behaviors, reshaping how professional services firms leverage internal knowledge.