Analysis · 30 September 2026 · 6 min read

Mustafa Suleyman's white collar AI prediction

Suleyman said most lawyer, accountant and marketing tasks would be automated within 18 months. What Anthropic's usage data shows seven months in.

In February 2026, Microsoft AI CEO Mustafa Suleyman told the Financial Times that most tasks done by lawyers, accountants, project managers and marketing staff would be "fully automated by an AI within the next 12 to 18 months". That puts the deadline between February and August 2027. Seven months in, Anthropic's usage data shows AI doing about a sixth of lawyers' tasks and about a third of accountants' tasks in real work, well short of most.

What he said

The full quote, as reported from the FT interview: "white collar work where you're sitting down at a computer, either being a lawyer or an accountant or a project manager or a marketing person, most of those tasks will be fully automated by an AI within the next 12 to 18 months."

The word to notice is tasks. He did not say those jobs would disappear. He described the people in them moving toward oversight and strategy while AI does the routine work.

It is the same shape as Anthropic CEO Dario Amodei's January prediction that AI would do most of what software engineers do within six to twelve months, extended to office work in general.

How much of these jobs AI is doing now

Anthropic's March 2026 labor market study measured, for each occupation, the share of tasks already showing up in automated, work related use of its models.

Observed exposure from Anthropic's labor market impacts data. Task times from its June 2026 Economic Index release, which estimates how long the tasks people bring to Claude would take without and with AI. Project managers are not broken out in either dataset.

Two things stand out. Marketing research work is already above half, the only one of Suleyman's four jobs that is. And the tasks that are handed over shrink from hours to minutes, which is why his timeline does not sound absurd to the people building these systems.

Why most is still far away

A task counted in that table is one someone actually handed to a model. Most of what a lawyer or an accountant is paid for sits outside it: client judgment, negotiation, signing off on work someone will rely on, and being accountable when it is wrong. None of that is measured by a benchmark, and none of it moves at the speed of a model release.

The capability is moving fast. On GDPval, a benchmark built from real professional deliverables across 44 occupations, Anthropic's newest models now lead the independent leaderboard. But Anthropic's own economists found no systematic rise in unemployment among the most exposed workers since late 2022. The first measurable effect has been on hiring at the entry level, not on the people already in the jobs.

Where this leaves the prediction

Measured against his words, most tasks fully automated, the jobs he named sit between a sixth and two thirds of the way there on Anthropic's measure, with between four and eleven months left. It would take a much faster jump than anything seen so far this year.

Measured against the direction, he is right. The share of these jobs that AI handles is rising with every model release, and the work that goes first is the same everywhere: high volume, well specified and easy to check.

Our index scores for the four jobs he named: Lawyers (42), Accountants and Auditors (54), Project Management Specialists (43) and Marketing Managers (48).

Your own share

The figures above are averages for each occupation. Two accountants with different weeks have very different exposure. The assessment scores your own mix of tasks in about two minutes, free and without an account.

Figures come from the AI Job Risk Index and were current when this was published. Scores change as new research is ingested, so the index is always the live version. See how scoring works. Informational guidance based on published research, not professional career or financial advice.