Routine task intensity (RTI) is a key metric in labor economics research for predicting automation vulnerability. It measures the degree to which a job's work consists of tasks that follow explicit, codifiable rules — as opposed to tasks requiring judgment in novel situations, social interaction, or unstructured physical manipulation.
Economists classify tasks along two dimensions: cognitive vs. manual, and routine vs. non-routine. This creates four quadrants: routine cognitive (data processing, record keeping), routine manual (assembly line, sorting), non-routine cognitive (analysis, writing, strategy), and non-routine manual (plumbing, physical therapy). Automation has historically progressed from high-RTI roles outward.
A high RTI score means most of the job's output can be produced by following a fixed decision tree. These roles are highly susceptible to both software automation (for routine cognitive tasks) and robotics (for routine manual tasks). A low RTI score signals work that defies complete procedure capture — the hallmark of safe-zone tasks.
RTI is not just a binary classification — it applies at the task level within roles. Even in creative or social roles, there are sub-tasks with high RTI (standardized reports, data formatting, templated responses) that AI can absorb, freeing the worker to focus on the low-RTI tasks that define the role's value.