Task decomposition is the methodological foundation of modern AI risk assessment. Rather than asking "will this job be automated?" — which is almost always an oversimplification — task decomposition breaks the role into its individual task components and asks which specific tasks are automatable, which are complemented by AI, and which remain fully human.
A typical job contains 10–30 distinct task types, varying widely in automation potential. A financial analyst's tasks might include: pulling data from databases (90% automatable), formatting standard reports (85% automatable), building financial models (60% automatable), interpreting model outputs in business context (30% automatable), and advising clients on strategy (10% automatable). The overall risk score is a weighted aggregate of these task-level assessments.
Task decomposition is more actionable than role-level analysis because it reveals exactly where to focus adaptation efforts. Instead of deciding whether to leave an entire profession, a worker can identify which tasks they should migrate toward (low automation risk, high value) and which they should learn to delegate to AI (high automation risk, commoditizing).
The methodology also reveals that two people with the same job title can face very different risk profiles depending on how their daily task portfolio is actually distributed. A senior accountant who spends 80% of their time on client advisory and 20% on preparation has much lower risk than a junior accountant with the opposite distribution.