Automation Curves: Where AGI Lands First, Hardest, and Last
Why AGI's effect on different jobs is uneven, the shape of typical automation curves, and what sectors are reshaping first.
Executive summary
Automation is uneven by design — it tracks task structure, training-data availability, and oversight requirements. Customer support, copywriting, data work, and routine code are reshaping first. Skilled trades, complex care, and trust-critical roles are reshaping much more slowly.
Key concepts
- Task structure
- Training-data availability
- Oversight requirements
- Regulatory pace
- Wage and demand response
What predicts speed of change
Three factors: how bounded and repeatable the work is; how much training data describes it; and how much oversight or trust the work requires. Bounded, well-documented, low-oversight tasks change fastest.
Sectors reshaping first
Customer support, content marketing, basic analytics, paralegal work, and translation are visibly reshaping in 2026.
Sectors reshaping slowly
Skilled trades, complex clinical care, hands-on engineering, trust-critical legal and financial advisory, and senior management.
Wage and demand response
Where automation makes a service cheaper, demand often expands. Healthcare and software both show this pattern repeatedly.
Key takeaways
- 01Automation tracks task structure, not job titles.
- 02Trust and oversight slow change in regulated sectors.
- 03Some markets expand as services become cheaper.
- 04Plan for a decade of uneven transition, not an overnight event.
Frequently asked questions
Will all knowledge work eventually be automated?
Highly uncertain. The credible cases for human-led judgment and relationship work remaining valuable extend well past the next decade.
What about embodied work?
Robotics is improving but is well behind cognitive AI on cost and reliability for unstructured environments.
Further reading
Related hubs
How AGI is reshaping knowledge work, automation curves, new roles, and organisational design — for individuals and organisations.
What changes for analysts, lawyers, engineers, designers, and other knowledge workers as AGI absorbs an increasing share of their tasks.
The categories of work that did not exist five years ago and are becoming significant — AI engineers, evaluators, prompt designers, AI safety roles, and more.