Methodology
How your assessment works
Fifteen answers about your work become five scores and an AI exposure band, using fixed rules. The result helps you think about where your work could change and what to explore next. It does not predict job loss. Here are the weights, thresholds and limits behind it.
The framework
Your answers feed five dimensions scored on a 0 to 100 scale. AI exposure is the headline index: it summarizes characteristics such as digital work, repetition, judgment and physical presence. A score of 70 does not mean a 70% chance of losing your job, or that 70% of your tasks can be automated.
Inputs · from your answers
- Digital intensity
- Task repetitiveness
- Rule-based work
- Information processing
- Content creation
- Data analysis
- Human judgment
- Interpersonal interaction
- Physical presence
- Accountability
- AI adoption
Dimensions · 0–100 each
- AI exposure10 inputs
- Automation exposure6 inputs
- AI augmentation potential6 inputs
- Human resilience4 inputs
- Transition readiness1 input
Result
- AI exposure · 0–100
Not shown above: expected outcome of a 2× productivity gain (feeds AI augmentation potential and Transition readiness) and willingness to learn or change role (feeds Transition readiness). These answers are listed with their weights below.
What moves each score
Each dimension combines a fixed set of answers, each with its own weight and direction. Bars extend to the right when a higher answer raises the score and to the left when it lowers it.
- Amber bar to the right
- Raises the score
- Neutral bar to the left
- Lowers the score
AI exposure
How much of your current work consists of tasks AI systems can already perform or materially assist with. Computer-based, repetitive, procedural, document-heavy and analytical work raises it. Physical presence, judgment in ambiguous situations, relationships and accountability lower it.
- Computer-based workraises AI exposure, 16% of the weight.
- Task repetitivenessraises AI exposure, 13% of the weight.
- Rule-based proceduresraises AI exposure, 12% of the weight.
- Information processing and analysisraises AI exposure, 11% of the weight.
- Document and content creationraises AI exposure, 10% of the weight.
- Physical presence and manual worklowers AI exposure, 16% of the weight.
- Human judgment in ambiguous situationslowers AI exposure, 9% of the weight.
- Relationships, persuasion and empathylowers AI exposure, 8% of the weight.
- Accountability for consequential decisionslowers AI exposure, 5% of the weight.
Automation exposure
The share of your work that could be handed to a system outright rather than assisted. Weighted toward repetitive, rule-based and computer-based tasks, and reduced where judgment, physical presence or accountability keep a person in the loop.
- Task repetitivenessraises Automation exposure, 24% of the weight.
- Rule-based proceduresraises Automation exposure, 20% of the weight.
- Computer-based workraises Automation exposure, 16% of the weight.
- Physical presence and manual worklowers Automation exposure, 20% of the weight.
- Human judgment in ambiguous situationslowers Automation exposure, 13% of the weight.
- Accountability for consequential decisionslowers Automation exposure, 7% of the weight.
AI augmentation potential
How much AI tools could raise your output without removing your role. Rises with computer-based, analytical and document-heavy work, with current AI use, and with an expectation that a productivity gain would make you more valuable. Dampened where the work is mostly physical.
- Computer-based workraises AI augmentation potential, 24% of the weight.
- Information processing and analysisraises AI augmentation potential, 16% of the weight.
- Document and content creationraises AI augmentation potential, 16% of the weight.
- Current AI adoptionraises AI augmentation potential, 14% of the weight.
- Expected outcome of a 2× productivity gainraises AI augmentation potential, 14% of the weight.
- Physical presence and manual worklowers AI augmentation potential, 16% of the weight.
Human resilience
How much of your value rests on things AI does not replace: being present, relationships and persuasion, judgment in ambiguous situations, and being the person accountable when a decision matters.
- Relationships, persuasion and empathyraises Human resilience, 28% of the weight.
- Human judgment in ambiguous situationsraises Human resilience, 28% of the weight.
- Accountability for consequential decisionsraises Human resilience, 22% of the weight.
- Physical presence and manual workraises Human resilience, 22% of the weight.
Transition readiness
How prepared you are to move with the change. Willingness to learn new tools or change role carries the most weight, followed by current AI use and openness to changing responsibilities.
- Willingness to learn or change roleraises Transition readiness, 50% of the weight.
- Current AI adoptionraises Transition readiness, 27% of the weight.
- Expected outcome of a 2× productivity gainraises Transition readiness, 23% of the weight.
The exposure bands
AI exposure is grouped into four bands using fixed thresholds. These are labels within this framework, not measured risk categories or a ranking against other workers. The thresholds do not move with other people’s results.
Low
0–34
Moderate
35–54
High
55–74
Very high
75–100
The band applies to the AI exposure index. The other dimensions use the same number range but describe different things. A readiness score of 70 does not cancel out an exposure score of 70; neither number is a probability.
Prepare without a countdown
The assessment cannot establish when your job will change. Timing depends on your employer, available tools, costs, regulation and how the work is organized.
Use your result to choose a small experiment. For repetitive digital work, test an approved tool on non-confidential sample material and check the output. For work centered on judgment or relationships, explore how AI might help with preparation while you keep responsibility for the result.
A 90-day plan gives you a schedule for your own actions. It is not a forecast of AI adoption or a deadline for changing careers.
About time windows in earlier results
Earlier versions attached year ranges to exposure bands. This legacy heuristic starts with the exposure band and adjusts it using reported AI use, repetition, defined procedures, physical presence and judgment. The stored field remains for compatibility with existing assessments and plans.
Those ranges were not calibrated against observed job changes. They are not validated timelines and should not be used to time a resignation, retraining decision or other consequential career move. Current preparation guidance uses the band without presenting a calendar forecast.
What this is not
A useful estimate depends on knowing its edges. These are the ones that matter most.
Your answers are self-reported.
The scores describe your work as you see it. We do not observe how you actually spend your time, and two people in the same role can answer differently and get different results. That is intended.
It is an estimation framework, not a scientific prediction model.
The weights encode a reasoned view of which task characteristics current AI systems handle well. They were not fitted to outcome data, and the result carries no confidence interval.
There is no occupation-level dataset behind it.
We do not look your job title up in a table of occupations. The title anchors the interpretation in the paid plan, but every number comes from your answers alone.
The narrative in the paid plan is interpretation, not measurement.
It is written by an AI model from your answers and scores. We validate its structure against a schema before showing it, but its judgments are the model's reading of your profile, and it can be wrong.
It is not a live measure of AI capability or hiring.
The framework uses fixed weights. It does not query current job listings, observe your employer or update itself when a new AI tool launches. Changes in tools, adoption and regulation can make its assumptions less useful.
A note on the paid plan
Where the model is used, and where it is not.
Your free work snapshot and first step use fixed rules based on your assessment answers. They do not call an AI model.
New paid-report purchases are closed during the free beta. Existing buyers retain their AI-written reports. The explanation below applies to those paid reports; the public sample is fictional.
The paid Career Transition Plan is written by an AI model from Anthropic, working from your job title, industry, answers and computed scores. The model is used only for interpretation: suggesting tasks that may fit your profile, sketching plausible scenarios and proposing concrete next steps. Those task suggestions are inferred from your answers, not verified observations of your working day.
It never computes or adjusts a score. Your numbers are fixed the moment you finish the assessment, and regenerating a plan does not change them. Every plan is checked against a fixed structure before it is shown, and nothing that is not in that structure is rendered.
Your email address is never sent to the model provider, and we do not use your answers or your plan to train any model. The privacy page lists exactly what is stored and for how long.
See it applied to your job.
Fifteen questions, then a result you can read against everything on this page.
Free · ~3 minutes · No account required