Same Title, Scores of 55 and 31: Hours Decide AI Exposure

The short answer: both managers are competent people doing legitimate versions of the same job. One spends the week producing text and analysis, which rates high. The other spends it on hiring conversations, vendor negotiation, and events, which rate low. Average the task ratings without weighting by hours, as a title-based tool does, and you get 52: close to the first manager, twenty-one points off the second, and describing a week nobody has.

The first manager

The first spends the week producing text: research synthesis, forecasts, budget models, campaign material. On the task ratings those hours land high. Compiling product offering lists rates 90 on exposure. Analysing market research rates 80. Those tasks take most of the week, so the weighted average sits high, at 55.

None of this makes the first manager worse at the job. It makes their week a set of text transformations, which is the category current models handle best.

The second manager

The second spends eleven hours a week on hiring and performance conversations, nine negotiating vendor contracts, and eight on events and trade shows. Those tasks rate 15, 20, and 30. Same task list, opposite end of it. The weighted average comes out at 31.

The second manager still writes. They still forecast. Those hours exist in their week, but they are a minority of it, and the average reflects the majority.

The part worth keeping

Average the nine task ratings without weighting by hours, which is what a title-based tool does, and you get 52. That lands within three points of the first manager and twenty-one points off the second.

An unweighted average describes a person who spends equal time on every task in their occupation. Nobody has ever had that week. The number looks precise and is precisely wrong for almost everybody, and it is wrong in a specific direction: it pulls people with unusual weeks toward the middle, which is exactly the people who most need to know they are unusual.

What the gap does and does not show

Neither number predicts anything about either person. Manager A is not on a two-year clock and Manager B is not safe. Both work at the same company, subject to the same adoption decisions, the same quality bar, and the same demand for marketing managers. The gap between exposure and outcome applies to both of them equally.

What the gap shows is how much signal a ranked list of occupations discards before it even starts. Twenty-four points of difference between two people with one title is the measurement itself, and a title-based tool throws it away in the first step.

Which way have your own hours moved?

Most people's weeks drift over a few years. A marketing manager who spent 2023 on campaign copy and now spends most of the week on agency management has moved from one column to the other without changing title. A manager whose team shrank and who now writes more of the copy personally has moved the other way. The score is built to see that drift.

So the question for your own week is not "what does my title score." It is: which way have my hours moved over the last three years, and which way do I want them to move over the next three? The free score gives you the current position. The direction is yours.

Frequently asked questions

Can two people with the same job title have different AI risk scores?

Yes, and usually do. The score weights each task's exposure rating by the hours a person reports on it. Two marketing managers with different weekly mixes can score 55 and 31 under the same title.

What does an unweighted title-based score get wrong?

It averages all of an occupation's task ratings equally, which describes someone who spends the same time on every task. For the two managers above it produces 52 for both, within three points of one and twenty-one points off the other.

Does a higher score mean the first manager is less skilled?

No. Both are competent people doing legitimate versions of the same job. The score reflects the composition of the week, not the quality of the work or the person.