There is no shortage of discussion right now about AI taking jobs. The question is not "Will AI Replace Me?" but rather “How much of my recent work actually required me specifically?”
To determine how at-risk your job could be, run the following audit:
Review your last 10 business days. Look at your calendar, sent emails, Slack or Teams messages, documents, tickets, spreadsheets, memos, reports, and any other place where your actual work shows up. Then tag each task, meeting, document, or deliverable with one of four labels:
T = Theater
C = Commodity
L = On the Line
D = Durable
The point is not to judge your entire job title. The point is to examine the specific units of work that fill your week.
T = Theater
Theater is work that exists because the organization performs it, not because it creates meaningful value. This includes things like:
Status meetings where nothing changes.
Decks nobody really reads.
Alignment calls that create no real alignment.
Reports that continue because someone asked for them 18 months ago. Reviews that exist mostly for political cover (large organizations often create rituals because rituals are visible, socially safe, and easy to point to as evidence that work is happening).
If a large portion of your week is theater, your job may look busy without being especially defensible.
C = Commodity Work
Commodity work is real work that produces value, but does not require you specifically.
Examples include:
Summarizing meetings.
Routing information.
Writing routine updates.
Applying known rules to familiar situations.
Creating first drafts where the structure is already obvious.
Turning meetings into next steps.
Coordinating decisions that have already been made.
This work matters. Companies run on it. But the risk is that it becomes less scarce. If someone else — or an AI tool guided by a decent prompt — could produce a similar output, then your personal claim on that work is weak. Just because a skill took years to develop does not mean the market will continue to reward it in the same way. A skill can be real and still become less scarce.
L = On the Line
“On the line” is the uncomfortable middle category.
This is work that still involves some human judgment, but is becoming increasingly repeatable or easier for AI to assist with.
Examples might include:
Pattern recognition in familiar situations.
Routine synthesis across known inputs.
Editorial judgment in an established format.
Relationship management based on known history. Work where a strong junior employee or AI-assisted worker could get 70% of the way there.
This category matters because it is where much of the anxiety lives. The work still feels skilled, but the boundary is moving. Some of it may become durable (see below). Some of it may become commodity work.
If much of your work falls into this bucket, your job may not be in immediate danger, but it is exposed. You need to actively move more of your time toward durable work.
D = Durable Work
Durable work is the work that still meaningfully needs a human.
This is not just “hard work.” Some hard work is highly automatable.
Durable work usually involves:
Judgement
Taste
Context
Ambiguity
Trust
Reading a room.
Reframing the problem.
Understanding what is not being said.
Making a call when the data is incomplete.
Seeing that the stated problem is not the real problem.
AI is very good at answering questions when the question is already well-framed. Durable human value often appears before that — when someone realizes the question itself is wrong.
For examples:
A customer asks for a feature, but the feature is not what they actually need.
A strategy looks clean on paper, but someone with experience can see it will fail in practice.
A project is moving quickly, but in the wrong direction.
That kind of work is harder to automate because it depends on context, judgment, and social intelligence.
How to Read Your Results
Once you tag your work, add up the categories.
The most important number is probably:
Theater + Commodity = the portion of your week on thin ice. It means that part of your week is the least defensible in an AI-enabled workplace.
A rough way to interpret the results:
If most of your work is T, you may be busy but not creating much enduring value.
If most of your work is C, you are doing useful work, but your role may be vulnerable to automation, delegation, or restructuring.
If much of your work is L, you are in the transition zone. You need to consciously move toward higher-judgment work.
If a meaningful portion of your work is D, you likely have a stronger foundation — especially if your organization recognizes and values that judgment.
What to Do After the Audit
Once you understand where your time is going, you can make better decisions.
First, stop volunteering for theater. Cancel or shorten unnecessary meetings. Reduce low-value reporting. Turn rituals into async updates where possible. Start with the low-risk items nobody will miss.
Second, use AI to compress commodity work, but do not reinvest all the saved time into producing more commodity work. If AI helps you write updates faster, the answer is not necessarily to write twice as many updates.
Third, move your recovered time toward durable work. Seek out messy projects, ambiguous customer problems, unclear strategic questions, and situations where judgment matters.
Fourth, build a private record of your judgment calls. Each week, write down one decision or intervention where your judgment changed the outcome. Over time, this becomes a portfolio of your durable value.
Fifth, make your durable work visible enough to be valued. You do not need to turn every instinct into a process document, but you do need to communicate outcomes.
Finally, be honest about your role. Some roles are structurally full of theater and commodity work. If there is no realistic path to more durable work, the best move may be to start looking for a role with more ambiguity, judgment, and direct exposure to real business problems.