McKinsey’s Skill Change Index maps 10,500 skills by automation exposure. The pattern is clear: technical execution skills (SQL, software dev, accounting) face the highest automation risk. Human judgment skills (resilience, empathy, leadership) face the lowest. The skills that survive are the ones machines can’t fake.
Skill Change Index — Automation Exposure
10,500
Skills indexed by automation exposure
<15%
Change index for resilience, empathy, leadership
2x
Gap between most and least exposed skills
The Skill Map
McKinsey Global Institute’s Skill Change Index (May 2026) ranks ~10,500 skills on a 0-100 scale by how much automation reshapes them. The top 100 skills cluster into a revealing pattern when plotted by automation exposure.
Most Exposed to Automation (Index 30+)
SQL / Programming — highest exposure. The skill AI replicates most directly.
Accounting / Invoicing — structured, rule-based, highly automatable.
Software Development — Block’s BuilderBot writes 15% of production code. The trend is accelerating.
Quality Control — inspection, testing, verification. AI does it faster and doesn’t fatigue.
Mid-Range Exposure (Index 20-30)
Research / Problem Solving / Analytical Skills — AI augments but can’t fully replace. The human frames the question; AI processes the data.
Detail Orientation / Collaboration — partially automatable. AI handles the detail; humans handle the collaboration.
Least Exposed (Index <20)
Leadership / Innovation — setting direction, making judgment calls under uncertainty. The Builder-PM’s territory.
Empathy / Influencing — human connection. AI can simulate it. Humans can feel it. The difference matters.
Resilience — the lowest automation exposure of any skill. The ability to absorb setbacks and adapt. No model replicates this.
The key insight: The skills most exposed to automation are the ones that can be described as instructions. SQL is instructions. Accounting is instructions. QA is instructions. The skills least exposed — resilience, empathy, leadership — cannot be reduced to instructions. That is the dividing line.
The Structural Read
This is the Harness Society thesis in data form. The principal/operator split runs exactly along this chart’s curve:
OPERATORS = HIGH EXPOSURE
SQL, accounting, software dev, QA — the execution layer. These skills are being absorbed by AI tools. The people whose primary value is executing instructions face the highest displacement risk.
PRINCIPALS = LOW EXPOSURE
Leadership, empathy, resilience, influencing — the judgment layer. These skills are what remains when execution automates. The people whose primary value is deciding what to do, not how to do it, face the lowest risk.
THE MIDDLE = CONTESTED
Research, problem solving, analytical skills — the skills that can go either way. AI augments them today. Whether the human retains control depends on whether they’re framing the problem or just processing the data. Framing survives. Processing doesn’t.
The Bottom Line
McKinsey indexed 10,500 skills and the pattern is binary: if a skill can be described as a set of instructions, AI is coming for it. If it requires judgment, connection, or the ability to absorb uncertainty — it survives. The career advice that follows is equally binary: move up the chart. From SQL to problem-framing. From accounting to strategic finance. From software development to bet selection. The skills at the bottom of the automation curve are the ones the Builder-PM Manifesto calls authorship — wanting, choosing, answering for outcomes. That is the Accountability Floor. Everything below it automates.
Business Engineer Framework
The Harness Society — The Principal/Operator Split in Data
McKinsey’s Skill Change Index is the empirical confirmation of the Harness Society’s central thesis: AI compresses the operator layer and concentrates value in the principal layer.
Source: McKinsey Global Institute — “Agents, Robots, and Us: How AI Reshapes Work and Skills in Europe,” Exhibit 12, May 2026