Two people in the same sector, even the same team, can face completely different AI exposure. Industry averages hide that. Here is why the risk sits in the role, not the label.
Ask about AI risk by industry and you will get a single number back: a sector average, a headline, a place on a leaderboard. It feels precise, and it is almost useless for an individual. Two people in the same industry, sometimes the same team, can face completely different AI exposure. The risk does not live in the sector. It lives in the role, and the industry average blends very different roles into one figure that describes none of them well.
AI compresses tasks, not sectors
The reason exposure varies inside an industry is that AI acts on tasks, not on job titles or sectors. Every industry contains a spread of work: routine production at one end, accountable judgment at the other. Finance holds both the analyst who reconciles and formats reports all week and the one who sits in the room where capital is committed. Law holds both routine document review and the partner who owns the client relationship and the risk. Same field, opposite exposure, because the underlying task mix is opposite.
- Exposed within a sector: high-volume production, data handling, formatting, first-draft content, and routine review.
- Durable within the same sector: decisions, client ownership, licensed or physical work, and named accountability.
- The industry average is just the blend of those two, weighted by how many of each the sector employs.
Why the average misleads
A sector average is a real number, but it answers a question almost nobody is actually asking. Nobody works as an average. You work in a specific role, with a specific week, and your exposure is set by the composition of that week, not by the sector you happen to sit in. A high-exposure industry still holds durable, judgment-heavy roles that are barely touched. A low-exposure industry still holds back-office desks doing routine work that compresses. Reading the average and stopping there is how people either panic without cause or feel safe without reason.
How to read the risk in the role
The reliable read is structural and role-level. Break your work into its parts and look at how much is routine production a model could already do, where your authority and accountability sit, and how much of your output could be generated without you. That is exactly what the Compression Exposure Index measures: the specific structural conditions of a role, which is why two people with the same title can read very differently. The methodology behind it is deterministic, so the reading reflects structure rather than a model's opinion.
Using the industry view correctly
The industry lens is still useful, as long as you treat it as the entry point and not the answer. It tells you the shape of a field, where compression tends to concentrate, and which neighbouring roles are more or less durable. The AI career risk by industry directory is built for exactly that: start at the sector, see the spread, then drill into the specific roles. A finance professional, for instance, can open the finance industry hub and compare the routine desks against the judgment-heavy ones instead of accepting a single sector figure.
From there the move is to read your own role, not the label above it. A financial analyst, a paralegal, a nurse, and a software engineer each carry a distinct structural profile, and even within those titles the exposure depends on how the individual week is actually spent. The sector average is a starting map. Your role is the territory, and the territory is what determines where you stand.
Why the spread is widest where it matters
The sectors people worry about most tend to have the widest internal spread, which is exactly why the average is least useful there. Take a knowledge-work field: it can hold a large tier of routine production, the reporting, the drafting, the reconciling, sitting right next to a smaller tier of high-authority, accountable work. Average those together and you get a middling number that flatters the exposed tier and understates the durable one. The people in each tier experience the field completely differently, and a single figure erases that difference precisely where the stakes are highest.
It also changes over time in a way an average hides. As AI absorbs the routine tier, the composition of the sector shifts: the exposed desks thin, the durable ones become a larger share of what remains, and the average creeps upward even though no individual role changed. Reading at the role level is the only way to see that movement, because it separates the desk that is compressing from the desk that is quietly becoming more valuable. For anyone planning a career inside a shifting sector, that distinction is the whole game: it tells you whether to double down where you are or to move toward the tier that is gaining ground.
Run your structural assessment
The AI Career Index is free to take. Get your structural snapshot in 10 minutes, including your durability score, archetype, and projection.
Take the Free Assessment