The dominant narrative around AI and work is still framed interms of jobs

- jobs lost

- jobs created

- jobs augmented

This framing is intuitive, but ultimately misleading.

AI does not operate at the level of jobs. It operates at the level of activities.

Every job is a bundle of tasks: some routine, some judgement‑based,some relational, some creative. AI does not replace the job as a whole; it selectively targets individual activities within it. The result is not a clean substitution of humans with machines, but a progressive decomposition of work.

‍

Some activities are automated. Others are accelerated or augmented. Some become more valuable because they are uniquely human.

And entirely new activities emerge as AI reshapes workflows.

This is the real story, not job loss, but task‑level transformation.

Why the Job-Level Lens Fails

When organisations plan for AI at the job level, they missthe granularity that actually determines impact. This leads to predictable errors:

• Overestimating displacement - Because they assume whole jobs disappear rather than specific activities.
• Underestimating redesign - Because they overlook how remaining tasks can be recombined into new roles.
• Struggling to translate AI adoption into workforce strategy - Because job descriptions are too coarse to guide reskilling, hiring, or organisational design.

The job is a convenient administrative construct, not ameaningful unit of analysis for AI transformation.

The Real Unit of Analysis: Activities

Once work is decomposed into activities, a different set ofstrategic questions emerges, questions that leaders rarely ask today:

• Which activities create differentiated value? These should be protected, enhanced, or elevated
• Which activities can be automated, accelerated, or outsourced? These become candidates for AI-driven redesign.
• How should the remaining activities be recombined into new roles or workflows? This is where innovation in organisational design happens.

This shift, from jobs to activities, is not a semantic nuance.

It is a structural rethinking of how work is understood, measured, and transformed.

What Changes When Organisations Shift to an Activity Lens

A task-level understanding of work reshapes multipledomains:

1. Workforce Planning - Instead of forecasting job supply and demand, organisationsmodel activity portfolios, identifying where human capability isessential and where AI can take over.

2. Learning and Development - Reskilling becomes targeted and modular. Employees learn thecapabilities needed for specific activities, not broad job categories.

3. Organisational Design - Roles become more fluid, dynamic, and continuouslyreassembled as AI reshapes activity sets.

4. Performance Management - Evaluation shifts from job outputs to activity-levelcontribution, collaboration, and value creation.

5. Talent Strategy - Hiring focuses on capabilities, not job titles, especially capabilities that complement AI rather than compete with it.

‍

AI Is Not Eliminating Work; It Is Illuminating It

AI is not destroying work. It is forcing organisations tounderstand work at a level they have historically ignored.

For decades, jobs have been treated as monolithic units, stable, fixed, and neatly defined. AI exposes the fiction. It reveals theunderlying architecture of work: the activities, micro‑tasks, and decisionpoints that actually create value.

‍

Organisations that embrace this decomposition will notsimply adapt to AI. They will redesign work more intelligently than their competitors.

‍

Those that cling to job-level thinking will misdiagnoserisk, misallocate investment, and miss the opportunity to build a moreadaptive, human‑centred workforce.