How to tell if your job is AI-vulnerable, in five honest questions.

Most "is your job at risk from AI" articles online are either fear-bait or marketing. We wanted to write the version we wished existed when our own team members started asking the question privately, in the hallway, after the second espresso of the morning.

The honest answer is: it depends on five things. None of them are about whether you're "good at your job." All of them are about which parts of your job AI can do well today.

Question 1: How much of your week is spent on patterned, repeatable output?

If your typical week is dominated by tasks that have predictable inputs and outputs — writing similar reports, processing similar tickets, summarizing similar meetings, producing similar briefs — you are more exposed than average. This is the category AI compresses fastest.

Conversely, if your week is dominated by genuinely novel problems, ambiguous stakeholder decisions, or in-person work, you are less exposed. The pattern that the data keeps showing in 2026 is that routine reduces over time and judgment doesn't.

Question 2: Could a smart intern with AI tools do your job at 70% quality?

This is a question we got from a manager who quietly automated half of his analyst team's deliverables and then froze hiring. He said: "I'm not firing anyone. I'm just not replacing the two who left."

If the answer is "yes, with the right AI tools, a sharp generalist could approximate my output," you are in the medium-exposure category. The job survives, but the per-role headcount slowly thins.

Question 3: How much of your value is relationship-based?

This is the one most people underweight. Account executives, partner managers, senior consultants, and people-managers all have one thing in common in 2026: their relationships are the moat. AI compresses the work around those relationships, but it doesn't replace the relationships themselves.

If significant portions of your value come from "people trust me, specifically," your exposure is lower than the surface-level analysis suggests.

Question 4: Do you currently use AI tools daily?

This question is unfair, but the data is clear: the people using AI are not the ones losing their jobs. They are the ones absorbing the productivity of the people who aren't.

From the manager's perspective, the calculus is straightforward: when you ask the team to do more next quarter, you keep the people who can deliver more without burning out, and you let attrition handle the rest. AI fluency is the silent prerequisite for "can deliver more."

Question 5: How long would it take you to get back to AI-fluent if you started today?

This is the single most useful question on the list, because it's the only one you can change. The labor economists we interviewed converged on the same window: 6 to 18 weeks of consistent practice closes the visible productivity gap for most knowledge workers in 2026.

The work is not technical. It is habitual. The professionals who close the gap don't take 80-hour bootcamps. They spend 20 minutes a day for three months on the actual work they do, using the tools.

The honest scoreboard

If we had to summarize the five questions into a rough scoreboard, it looks like this:

• High Patterned output + smart-intern-replaceable + no AI use today + no relationship moat = highest exposure. Catch-up window urgent.
• Med Mixed output + partial AI use + medium relationship value = moderate exposure. Productivity expectations rising; act this quarter.
• Low Novel work + strong relationships + already AI-fluent = low exposure. Maintenance mode; keep learning at your own pace.

What to actually do about it

If you scored "medium" or "high" on the informal scoreboard above, the answer isn't to panic. It's to spend twenty minutes a day, for ten to twelve weeks, getting genuinely fluent at the small set of AI tools your job actually needs. That is the entire intervention.

The people who close the gap aren't smarter than you. They're earlier. And the cost of "earlier" gets steeper every quarter.