Which Parts of Any Role Should Stay Human — and Which Can an Agent Handle?

Role decomposition breaks a job title into its 8–15 discrete tasks and classifies each as Human (requires judgment, empathy, or accountability), Agent (AI handles reliably), or Hybrid (AI executes volume, human oversees exceptions). It's the first step in designing a hybrid workforce — you can't optimize a role without understanding its task-level composition first.

Classifications are based on four factors: (1) Task repetitiveness — structured, rule-based tasks favor Agent; novel, judgment-heavy tasks favor Human. (2) Error tolerance — low-stakes, auto-correctable errors support Agent; high-stakes, hard-to-reverse decisions require Human. (3) Data availability — AI needs structured digital inputs; unstructured or relationship-context data requires Human. (4) Current AI tool capability — assessed based on real tools available as of April 2026. All classifications are estimates labeled as PeopleStackHub.ai assessments.

Agent-ready percentage is the proportion of a role's weekly hours that could be handled by AI agent tools — including both fully Agent tasks and the AI portion of Hybrid tasks (estimated at 65% of hybrid task hours). A Customer Support Rep at 62% agent-ready means roughly 62% of their time goes to tasks where AI meaningfully reduces human hours. It does not mean eliminating 62% of headcount — a well-designed hybrid stack typically reduces headcount by 35–50% for that function while increasing overall output per McKinsey Global Institute 2025.

Yes. Enter any role title and optionally paste a job description. For roles not in the pre-seeded dataset, the tool uses AI analysis to generate a task decomposition. Pre-seeded roles use BLS O*NET task data for higher baseline accuracy. Custom role analysis is labeled as "AI-generated estimate" and should be validated against your specific context. If a role cannot be reliably decomposed, the tool will say so and suggest custom analysis.

Start by decomposing every role in a target function (e.g., Customer Support, Finance, Marketing). Stack the agent-ready percentages to get a function-level automation opportunity. Then use the 3-phase implementation plan to sequence investments — typically starting with high-confidence Agent tasks in Phase 1, adding Hybrid configurations in Phase 2, and optimizing the full hybrid stack in Phase 3. For full-organization analysis with cost modeling and hybrid blueprints, use the Workforce Design Agent at /design.