Skip to content

    Research analysis

    Analysis: automation removes parts of roles, not roles

    Author

    AI Cubed

    Published

    August 9, 2026

    Read time

    7 min

    McKinsey Global Institute's analysis of work activities found that for around 60% of jobs, at least 30% of the constituent activities could be automated with technology already available — while fewer than 5% of jobs can be automated entirely. Those are their figures.

    Our analysis concerns the operational consequence: if automation lands on activities rather than roles, then any plan built around job titles will target the wrong things and produce the wrong forecast.

    The distinction that changes the plan

    Two numbers sit in this research and they pull in opposite directions. The first — 60% of jobs with at least 30% automatable activity — says the opportunity is broad. The second — under 5% of jobs fully automatable — says it is shallow at the role level. Taken together they describe an operation where almost everyone gets lighter and almost nobody gets replaced.

    That is a far more useful planning input than a headline about job losses, and it is the version we work from.

    Our reading: scope by activity

    When a client asks which roles we would automate, we redirect the question. The unit of work is the activity: intake, verification, scheduling, data entry, reconciliation, reporting, follow-up. Each of those has a frequency, a rule set, an input quality, and an owner — and those four properties, not the job title attached to them, determine whether automation will hold.

    • High frequency, stable rules, clean inputs: automate now.
    • High frequency, unstable rules: fix the process first, then automate.
    • Low frequency, high judgement: leave it with a person and give them better information.
    • High frequency, messy inputs: fix the data, or you will automate the mess faster.

    What happens to the 30%

    In the operations we run, the reclaimed 30% is almost never the interesting part of anyone's job. It is form-filling, copying, chasing and checking. Removing it tends to raise the quality of the remaining 70% — a scheduler who no longer re-keys appointments spends that time on the exceptions that actually cost money when they go wrong.

    Automation removes portions of a role, not the role itself.Our summary of the McKinsey Global Institute finding

    How we plan against it

    1. Inventory activities across the affected function, not roles.
    2. Score each on frequency, rule stability, input quality, and ownership.
    3. Automate the top-scoring activities and leave the rest documented.
    4. Reforecast capacity — what the team can now take on — rather than cost removed.

    Frequently asked questions

    This page is AI Cubed's own analysis of independently published research. The underlying study belongs to its publisher, credited and linked in full below.

    Sources

    Start here

    See where your operation is losing time.

    Twenty minutes with an operator, not a salesperson. We'll name the one bottleneck costing you the most — and tell you whether it's worth fixing with software at all.

    Book your free 20-minute consult

    20 minutes · video call · no preparation needed