administrative assistant desk with laptop, notebook and coffee
    Back to Blog

    News and Commentary

    The New AI Job Data Points Straight at Admin Work

    Task by task, the research keeps landing on the same desk. Here is what I would do about it.

    Valentina Akpan, founder of Rellatech

    Valentina Akpan: Founder, Rellatech. Admin and technical virtual assistant for businesses and startups with teams.

    Last week I wrote about Microsoft cutting 4,800 jobs and the careful wording around AI. This week the more useful story arrived, which is the data on exactly who carries the risk.

    Yesterday, Forbes contributor Ron Schmelzer pulled the recent research together, and the pattern is hard to miss. Forbes reports AI is not clearing out whole job titles. It is sorting the labor market task by task, and the workers most exposed in the latest round of cuts were older white-collar staff, middle-skill office workers, entry-level analysts, and assistants.

    The numbers behind the pattern

    Brookings estimated in January that 37.1 million American workers sit in the top quartile of AI exposure. Most of them have room to adjust, whether that means transferable skills, savings, or a deep local job market. But 6.1 million people face high exposure with little capacity to adapt, and that group is concentrated in clerical and administrative occupations. Eighty-six percent of them are women.

    PwC's 2026 Global AI Jobs Barometer, built on more than one billion job ads across 27 countries, adds the other half of the picture. In highly exposed fields, junior roles are now seven times more likely to require senior skills such as judgment and leadership. Entry-level roles rewritten that way have grown 35 percent since 2019, while other entry-level roles have shrunk by 10 percent. The bottom rung of the ladder has moved up.

    Tasks, not titles

    The OECD's latest AI exposure measure explains why administrative work keeps showing up in these studies. The pressure lands on repeatable information processing, which means reporting, documentation, first drafts, and data entry. It lands far more lightly on work built around trust, relationships, judgment, and accountability.

    That matches what I see every week. The tasks going first are the ones I already automate for clients, like pulling reports, cleaning spreadsheets, drafting routine replies, and formatting documents. What has not gone anywhere is the layer above all of that. Someone still has to know which report actually matters, notice what the automation missed, and hold a calendar together when a client's week falls apart. Assistant work is not ending. The routine layer underneath it is thinning out, and the judgment layer is becoming the whole job.

    What this means if you run a business with a handful of people

    I would not hire a full-time role built mostly out of tasks that sit on the automation list, and I would not assume a software subscription covers the judgment either. The practical middle is administrative support that pairs automation on the routine work with an experienced human on everything that requires a decision. That is how I structure executive assistant engagements now.

    It is also exactly what Rellatech monthly retainers are built for. Businesses and startups with teams use them to keep the judgment layer covered while the routine layer keeps changing shape, without betting a full salary on a job description that may look different in a year. If you want to talk through what that could look like for your team, have a look at pricing or reach out and we can map it together.

    Contact Me

    Not ready for a call? Take the free operations assessment and I will send you a written breakdown within 24 hours.

    Share This Article

    Comments

    Join the conversation.

    No comments yet. Be the first.

    Leave a comment