How to Self-Check Your AI Compression Risk

You can run a five-question structural self-check that mirrors the AI Career Index methodology. It returns a directional read, not a calibrated score. The point of the self-check is to decide whether to take the full assessment seriously, not to substitute for it. Each question maps to a dimension that the deterministic engine actually scores; answering them honestly produces a useful proxy that takes about five minutes.

This article walks through the five questions, what each one is measuring, how to read the combined answers, and what to do with the read. The structural logic is the same logic the full index applies; the simplification is that you answer the questions yourself in plain language rather than running them through the deterministic scoring engine.

Question 1: routine cognitive load

What share of your typical working week is templated production work. Templated means the work follows a repeating pattern: same inputs, same transformation, same output shape. Examples include drafting weekly reports, producing recurring analyses, formatting documents, writing standard correspondence, and processing structured forms. Estimate the share as a percentage of the working week.

Read: a share above 60% indicates high routine load. A share between 30% and 60% indicates mid routine load. A share below 30% indicates low routine load. Higher is more exposed. The dimension matters because templated production work is the layer that AI tooling absorbs first, and the share of the week in that layer drives the structural risk.

Question 2: authority scope

How broad is your decision authority. Three levels work. Narrow: you execute decisions made by others, with limited scope to change the inputs, the method, or the output. Moderate: you propose decisions, scope the method, and own the recommendation, but the final call sits with someone else. Broad: you own named decisions with consequences, and the recommendation and the call sit with you.

Read: narrow authority is the exposed profile. Moderate authority is the mid-exposed profile. Broad authority is the insulated profile. Higher authority is more insulated. The dimension matters because authority-bearing work cannot be templated without removing the authority, which is a much bigger change than absorbing a templated task.

Question 3: accountability load

Whose name is on the output. If the output goes out under your name, with your judgment attached, your accountability load is high. If the output goes out under a process or a team name, with no single accountable person, your accountability load is low. The intermediate case is shared accountability: your name appears alongside others, but the final responsibility is collective.

Read: low accountability is the exposed profile. Shared accountability is the mid-exposed profile. Named, owned accountability is the insulated profile. Named accountability protects the role because the structural value of the role is the responsibility, not the production of the artefact. The artefact can be produced by tools; the responsibility cannot.

Question 4: relational anchor

Do you have a relational anchor in the work. Relational anchors include long-term clients, internal stakeholders who depend on you specifically, regulatory counterparts who know you by name, and team members who report to you. A role with a strong relational anchor cannot be re-tendered without losing the relationships; a role with a weak relational anchor can be re-tendered without consequence.

Read: no relational anchor is the exposed profile. Weak relational anchor (transactional internal stakeholders, generic client list) is the mid-exposed profile. Strong relational anchor (named long-term clients, named regulatory counterparts, named team members) is the insulated profile. The dimension matters because the relationship is the part of the work that AI tooling cannot replicate, and a strong relationship is a structural moat.

Question 5: tool dependence

How tool-dependent is the production of your work today. High tool dependence means the work is already produced through a small number of tools, with standard inputs and standard outputs. Low tool dependence means the work is produced through judgment, relationships, or physical practice that does not pass through a standard tool chain. The dimension is a proxy for how easily the existing workflow can be replaced by an AI-augmented workflow.

Read: high tool dependence is the exposed profile. Mixed tool dependence is the mid-exposed profile. Low tool dependence is the insulated profile. A role that already runs through a tool chain is one step away from running through an AI-augmented tool chain. A role that runs through judgment, relationships, or physical practice is structurally further from absorption.

Combining the five reads

Score each question one to three: one for exposed, two for mid, three for insulated. The sum runs from five to fifteen. A sum of five to eight is the exposed profile: the structural conditions are working against you on most dimensions. A sum of nine to eleven is the mid-exposed profile: the role is reshaping but not collapsing. A sum of twelve to fifteen is the insulated profile: the structural conditions are working in your favour and AI tooling is leverage rather than threat.

The combination matters more than any single answer. A high routine load is concerning if it is paired with narrow authority and low accountability, and is much less concerning if it is paired with broad authority and named accountability. The five dimensions are not independent; they form a structural profile, and the profile is the read.

Why this self-check works

The self-check works because the dimensions it measures are the same dimensions the full deterministic engine measures. The simplification is in the scoring (three buckets per question instead of fine-grained weights) and in the input (your honest estimate instead of structured questionnaire responses). The simplification trades calibration for speed. The directional read is reliable; the precise score is not.

The full version of the self-check is the AI Career Index assessment. The deterministic engine weights every dimension, runs a 24-month projection under drift versus intervention, and produces a structural archetype and the 6-month Acceleration Plan. The self-check gets you to the question of whether to take the full assessment; the full assessment gets you to the plan.

What to do with each profile

Exposed profile (sum five to eight): take the full assessment within the next two weeks. The structural conditions warrant immediate attention, and a structural plan is more cost-effective than a reactive plan after the next layoff or wage adjustment lands. Migration, climbing, and skill investment all return more when started early.

Mid-exposed profile (sum nine to eleven): take the full assessment within the next three months. The role is reshaping. The right intervention is structural migration inside the role, which the full assessment will scope. The pressure is real but not urgent.

Insulated profile (sum twelve to fifteen): take the full assessment at your convenience, primarily to confirm the read and to identify the next ceiling. The structural conditions are in your favour. The opportunity is to absorb AI tooling as leverage and use the reclaimed time to climb or to expand scope. The plan in this case is offensive, not defensive.

Limitations of the self-check

The self-check has three limitations to acknowledge. It depends on honest answers, which professionals close to a role often struggle with: routine load is systematically underestimated and authority scope is systematically overestimated. It does not capture industry-level exposure, which is one of the dimensions the full engine weights. It does not produce a 24-month projection or a structural archetype, which are two of the most useful outputs from the full assessment.

These limitations do not make the self-check wrong; they make it directional. A reading is directionally correct if it tells you whether to take the full assessment, where to start, and what to expect. That is the role of the self-check. The reading also gives a baseline you can compare against the full assessment result, which is useful for sanity-checking honesty.

From self-check to plan

The structural follow-up to the self-check is the full assessment. The AI Career Index Snapshot is free and returns the band, the headline score, the structural archetype, and the 24-month projection. The $49 Premium adds the full structural model, the projection comparison under drift versus intervention, and the 6-month Acceleration Plan calibrated to your weakest dimension.

The broader methodology that the self-check approximates is documented at the methodology page, which lays out the deterministic engine, the dimensions, the weights, and the derived metrics. Use the self-check as the entry point; use the full assessment as the structural plan; use the methodology page as the explanation of how the structural plan is generated.

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.