Is AI unbundling expertise?

2026-06-19 · Source: Thoughtworks Insights · Field: Business & Management — Corporate Strategy & Leadership, Human Resources & Workforce Development, Artificial Intelligence & Machine Learning · Depth: Intermediate, medium

Summary

An Anthropic study, published June 19, 2026, analyzed approximately 400,000 Claude Code sessions, revealing that AI is "unbundling expertise" by separating its traditional components: knowledge, reasoning, and transmission. While the study indicates that professionals like lawyers and accountants can achieve technical work success rates comparable to software engineers with AI assistance, the more significant observation is that effective AI utilization depends on a user's capacity to precisely transmit a coherent internal model to the AI. This "transmissibility" — the ability to make one's understanding legible to another intelligence — is emerging as an independent economic capability. Organizations that have historically valued execution over the explicit articulation of reasoning may find themselves underinvesting in this crucial skill, leading to a productivity bottleneck as AI amplifies clarity rather than inferring from ambiguity.

Key takeaway

For executives and HR leaders investing in AI fluency, recognize that success hinges on your team's capacity to transmit coherent internal models, not just their domain knowledge. Prioritize developing "transmissibility" through training in teaching, writing, and structured reasoning. Failing to cultivate this skill, which AI directly rewards, will create a significant productivity ceiling, regardless of tool adoption or prompting capabilities.

Key insights

AI unbundles expertise, making the precise transmission of internal models a distinct, economically valuable capability.

Principles

• Expertise is now decomposable into knowledge, reasoning, and transmission.
• Transmissibility, not just knowledge, multiplies AI system effectiveness.
• Rewarding explicability over execution is crucial for AI-driven productivity.

In practice

• Focus on making internal models structurally clear for AI systems.
• Cultivate skills like teaching, writing, and scientific training for transmissibility.
• Prioritize context selection over mere prompt engineering for AI success.

Topics

• AI Expertise Unbundling
• Knowledge Transmission
• Anthropic Claude Code
• Context Engineering
• Workforce Development
• Organizational Productivity

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Consultant, Executive

Related on AIssential

• AI Won't Replace You — But Someone Using AI Might — AI amplifies human productivity, making proficiency in its use a critical competitive advantage for professionals, rather than a replacement for human roles.
• Meta Layoffs: 'Success Isn't a Given', CEO Warns — Major tech companies are restructuring workforces and increasing AI investment, driven by AI's impact on productivity and organizational design.
• 🔮 Exponential View #562: Agents & the tedium frontier; AI in the statistics; robot insurance; Claude at war, hacking pigeons, AI dignosis++ — AI agents reduce delegation costs, acting as infrastructure for knowledge work and driving productivity gains.
• Shifting AI From Fear to Optimism: U.S. Department of Labor’s Taylor Stockton — AI is transforming nearly every job, necessitating widespread AI literacy and a shift to an opportunity-focused narrative.
• The AI Free Ride Is Ending: A Reckoning Is Coming for Software Engineering — The current AI-assisted development ecosystem relies on unsustainable investor subsidies, masking true costs and fostering dependency, leading to an inevitable pricing correction.
• Marc Andreessen on Builder Culture in the Age of AI — AI is a superpower expanding human capability and work, challenging traditional job roles and institutional narratives.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Thoughtworks Insights.