Mustafa Suleyman, head of Microsoft’s AI unit, said that AI could begin replacing human jobs within two years. The projection outlines a rapid shift for roles that primarily operate in front of screens and with standardized workflows.
Mustafa Suleyman predicted that it would take a maximum of two years for AI to replace human labor, at the same pace and with satisfactory results. According to his estimate, tasks in areas such as law, accounting, marketing, and project management are among the most vulnerable, especially when the work is dominated by routine processes performed in front of a screen.
In his timeline, positions based primarily on repeatable computer activity could be automated first, within a 12- to 18-month timeframe. Functions involving more complex coordination and interdependent decision-making chains would take somewhat longer to experience a comparable impact. If this scenario materializes, it would represent a significant disruption for millions of knowledge workers.
The scope he described is broad. In addition to the sectors mentioned, it also involves areas such as interpreting, translation, customer service, human resources, content writing, financial analysis, and data entry. The diversity of potentially affected functions raises immediate questions for companies with large back-office operations and for professionals whose output is structured and replicable.
Replacement versus augmentation: the debate on the real pace
However, Suleyman’s view is not unanimous, as several experts maintain that the most likely effect in the short term will be augmentation—that is, the automation of specific cognitive tasks within broader jobs—rather than the complete elimination of roles. They argue that most professions are composed of multiple activities, and only a few are easily replaceable.
Furthermore, they point out that the adoption of AI tools often generates new responsibilities related to supervision, integration, verification, and risk management. From this perspective, the transformation would be profound but gradual, involving a redefinition of work rather than immediate mass destruction.
The debate is amplified by other compelling projections. Anthropic CEO Dario Amodei stated that software engineering could become obsolete within 12 months, illustrating how the estimated timelines from industry leaders vary considerably. This divergence underscores that the uncertainty revolves more around the pace than the direction of change.
In practical terms, if Suleyman’s timeline proves accurate, companies will face a compressed schedule to adapt staffing levels, redesign workflows, and strengthen compliance frameworks related to AI use. Targeted retraining and the strategic integration of automated agents will move from experimental initiatives to immediate priorities.
For professionals, the implicit message is clear: the value of skills that combine industry-specific judgment, cross-functional coordination, and the ability to monitor outcomes generated by automated systems will increase. Beyond the exact timeframe, the conversation is no longer about whether AI will transform office work, but rather how—and how quickly—organizations will manage that transition.
