My Take on the McKinsey Report: Automation Is Coming for Tasks, Not Entire Professions

My Take on the McKinsey Report

The image above appears to be based on McKinsey’s Skill Change Index, highlighted in Agents, robots, and us: Skill partnerships in the age of AI and referenced again in Skills reset for the AI age.¹²

McKinsey’s broad message is clear: some skills are far more exposed to automation than others. Digital and information-processing skills sit higher on that exposure scale, while caring and people-centered skills sit lower.¹²

My take is that the chart is directionally right, but still too broad.

The biggest mistake people make is thinking AI will replace professions in one clean sweep. That is not what is happening. AI is stripping away routine layers of work inside almost every profession.

That is especially visible in software.

Routine coding, boilerplate generation, basic reporting, repetitive analysis, and standard workflows are getting cheaper fast. But architecture, systems thinking, production judgment, trade-off analysis, and accountability are not getting commoditized at the same rate.

So the real divide is not job versus job.

It is task versus task.

Tasks that are repetitive, rules-based, low-context, and easy to verify are more exposed. Tasks that are ambiguous, cross-functional, high-stakes, and expensive to get wrong are less exposed.

That is why I do not read this chart as “coding is doomed” or “some jobs are safe.” I read it as a warning that the market is starting to reward a different layer of value.

In software and other knowledge work, syntax alone is becoming cheaper. Judgment is becoming more expensive.

That is the shift.

The people who will do well are not just the ones who can produce output faster. They will be the ones who can frame the right problem, validate AI output, understand how systems fail, and make better decisions when the trade-offs are messy.

That is where the durable value is moving.

References