Role assessment
Analyze a role to see its recurring tasks, estimate where AI could help, and pick one task to test.
Example library
Open a fictional report to see what’s worth testing and where a person stays in charge.
Fictional example
Dispositions & Transactions
Role directory
06 EXAMPLESScoring methodology
Four ratings set the role estimate. Each is rated from 1 to 5. A lower rating means fewer barriers to automation.
Worked example
Start with ratings of 4 for complexity, 3 for decision impact, 2 for human skills, and 1 for physical activity.
For an individual task, AI feasibility alone sets the estimate. A 4 out of 5 equals 75%High.
02 / TASK AUTOMATION
Name the task you want to automate, define how to test it, and give the owner a six-week brief they can act on.
Draft saves in this browser. Use sanitized examples.
Frequently asked questions
A role is a bundle of tasks. A task score helps identify work to test with AI. Employment outcomes depend on far more than the ability to automate one task.
Capability changes unevenly. Recheck an estimate when your tools, inputs, permissions, or quality requirements change.
Start with work that repeats, uses structured inputs, and has a clear output you can verify. Examples include extracting fields, classifying requests, and drafting standard updates.
Tasks that rely on physical presence, relationship judgment, or accountability usually need more human involvement. AI can still assist with preparation and follow-up.
Live analysis uses your department, role title, and experience level to suggest common tasks and assign five-point factor ratings. The site turns the ratings into role and task estimates.
It summarizes how suitable a role or task may be for AI assistance under the model’s assumptions. It helps you choose what to test first.
A role title can conceal important differences in what people do. Check the suggested tasks against your real work, then use a supervised pilot to test quality and time saved.
Live reports are generated on request. Run a new report as your role changes, and use recent examples to validate any proposed pilot.
Pick one frequent task with low consequences if a draft is wrong. Set a baseline, a review point, and a stop rule before testing.
Practice defining good inputs, verifying evidence, judging exceptions, and owning decisions. Those skills carry across tools.
Start with a draft, checklist, or summary you can verify. Record the time spent reviewing and correcting it as part of the pilot.
Map recurring work with the people who do it. Agree on approved sources, accountable owners, and success criteria before broader rollout.
Use authorized data, minimize what you share, explain where AI is used, and put a responsible person in charge of decisions that affect people.
