What Is the Second Job of Knowledge Work?

The second job of knowledge work is the unnamed work of managing work: the prioritizing, coordinating, communicating, and improving that every knowledge worker performs alongside their professional specialty — without ever having been trained for it. The term comes from the Work Management Thesis, published by the Work Management Institute, which argues that this second job is becoming the defining human skill of the AI age.

The Job That Isn't on Your Business Card

Ask professionals what they do and they'll tell you about their first job — the one they trained for. A finance leader manages budgets and forecasts. A marketer runs campaigns. An engineer builds products.

Watch how they actually spend a Tuesday and you'll see the second job: reconciling three conflicting priorities before lunch, chasing a status update that should have arrived yesterday, translating a decision from one team into action items for another, rescuing a handoff that quietly stalled, and deciding — again — what has to slip so that something else can ship.

None of this appears in a job description. All of it determines whether the first job succeeds.

And it isn't a small slice of the day. Asana's Anatomy of Work Index, a study of more than 10,000 knowledge workers, found that 60% of a person's time at work goes to "work about work" — coordination, status-chasing, and searching for information — rather than the skilled work they were hired to do. The invisible job is the majority of the job.

Why Has No One Named It?

Because it's everywhere, it belongs to no one. The second job isn't a role, a department, or a line item. It's distributed across every knowledge worker in every function, which is exactly why organizations have never treated it as a competency. There's no degree in it. No onboarding module. No performance review category. The universal assumption is that people will simply pick it up through experience.

Some do. Most absorb the cost instead — as workflow debt, meeting overload, invisible work, and the low-grade coordination friction that everyone feels and no one owns.

Software vendors got closest to naming it. "Work about work" described the symptom well, but framed it purely as waste — friction for a platform to eliminate. That framing stops short of the more important truth: this labor is a skill. It can be done well or badly. Done badly, it produces the dysfunction above. Done well, it's the difference between a team that ships and a team that spins. And anything that can be done well or badly can be taught — which makes it a discipline, not just a nuisance.

That discipline is work management.

Why the Second Job Suddenly Matters

For decades, the second job could stay invisible because the first job was where human value lived. That's the assumption AI is breaking.

AI now drafts the campaign, analyzes the financials, writes the code, and builds the report — absorbing more of the first job every quarter. What it cannot absorb is the judgment core of the second job: deciding what work is worth doing, setting priorities among competing demands, designing workflows that span people and intelligent agents, and owning the outcome when it matters.

That's the inversion at the center of the Work Management Thesis: as AI takes over the first job, the second job becomes the first. The work is being automated; the management of work is being promoted. The skill nobody was trained for is becoming the skill that's left.

What to Do About Your Second Job

Three starting points:

Name it. Track one week of your time honestly and separate first-job work from second-job work. Most people discover the Anatomy of Work numbers are not an exaggeration. You can't improve a job you haven't admitted you have.

Treat it as a skill, not a tax. The second job has real practices — clarifying priorities, designing explicit handoffs, running decisions transparently — that can be learned deliberately instead of absorbed accidentally. That's the entire premise of work management as a discipline.

Let AI take the mechanical layer. Status compilation, deadline nudges, and report assembly should be automated. What you keep is the judgment layer — and that's the layer worth getting trained in, because it's the part of your job that AI makes more valuable, not less.

The second job has been hiding in plain sight for the entire history of knowledge work. Naming it is the first step. Building the discipline is the work ahead.

Read the full argument in the Work Management Thesis.