Everyday errands, handed to an agent

One page per errand: what the job really involves, where agents break, and what the exam checks. Not a feature list.

• Booking a tableAnd the errand that quietly races other peopleAI restaurant reservation tools today: what Google's AI Mode does, where a personal agent goes further, and what our exam measures when one books a table.
• Getting a subscription cancelledAnd what that asks of an agentAn app to cancel subscriptions has to get past a flow built to keep you paying. What that errand asks of a personal agent, and what our exam measures.
• Disputing a chargeAnd the decision an agent should not make for youHow to dispute a charge: report fraud to your card issuer, take a bad purchase to the merchant, the US deadlines, and what changes when an AI agent does it.
• Rearranging a family weekAnd who has to be toldWhat an AI family assistant like Cozi or Ohai does today, what a personal agent claims beyond it, and how our exam tests one that messages other parents.
• Claiming for a delayed flightAnd the offer an agent should bring back to youFlight delay compensation: what US, EU and Canadian rules pay, why a voucher can cost you the refund, and what an AI agent must not accept for you.
• Watching something liveAnd knowing when not to tell youNotification overload and notification fatigue: what focus modes already do, where they stop, and how our exam tests an agent that decides what reaches you.
• Organising work filesAnd the errand where finished is hardest to defineWhat an AI file organizer like AI File Sorter or Sparkle does, what a personal agent claims beyond it, and how our exam checks a folder an agent tidied.

Individual attempts at these errands, with every measurement listed separately, are on the recorded runs page. How the exam turns an errand into a task is on how the exam works.