The assistant has booked the aisle seat. The passenger wanted the window.

Neither choice was absurd. That is the point. The agent was asked to arrange a trip, and it quietly answered a question nobody had asked it: which small comfort matters more?

As agents move from answering to doing, this becomes a brand question. Every default is a little statement of intent. It tells a customer, often without words, what the organization thinks can be decided on their behalf.

The default has already spoken

The idea that a chosen default shapes behaviour is not new. In their 2003 paper, Richard Thaler and Cass Sunstein argued that choice settings cannot be neutral: somebody has to arrange them. A later meta-analysis of 58 studies found that defaults generally influence choices, but not evenly. They work differently when people read them as an endorsement or as the normal state of affairs.

That variation matters for brands. A default delivery slot may feel like help. A default insurance add-on may feel like a hand already in the customer’s coat pocket. The same efficient click can signal care or presumption.

An agent gives defaults more reach. It can choose a product, a time, a tone, a route, or a level of disclosure while the person is busy looking for their keys. The brand is no longer only the message on the receipt. It is the judgment inside the helpful shortcut.

BUREAU SYSTEM NOTE 19: The customer has been spared a decision. Please retain the receipt for later feelings.

A pause can be part of the promise

The hard part is not making an agent ask more questions. A cashier who asks for a life history before selling a pencil is not considerate. They are creating a queue with ambitions.

The hard part is recognising the choices that belong to the person. Google researchers reported in 2025 that language-model agents often guess an ambiguous intent instead of asking a clarifying question. Their work improved agents’ ability to recognise when clarification was needed (Chen and colleagues).

A separate 2025 study of an accessible UI agent made this concrete. Peng and colleagues designed an agent that stopped at consequential choice points rather than finishing the whole task by itself. In their evaluation, people completed more tasks and chose options that better matched their preferences than with the comparison agents. This is one study in a specific setting, not a universal rule. It is still a useful correction to the slogan that friction is always failure.

For a brand, the useful distinction is simple:

  • Proceed when the choice is reversible, routine, and clearly in the customer’s stated interest.
  • Show the default when speed helps but a reasonable person may want a different route.
  • Ask when the choice affects money, privacy, identity, safety, taste, or a promise the brand has made.

This is not a compliance checklist in a trench coat. It is a way to make the brand’s manners visible. A thoughtful interruption can say: we are here to help, not to complete you.

Make the pause deliberate

The conceptual discipline is intentionality: under faster systems, leaving a choice unspecified does not leave it open. It hands the choice to the system’s habits, incentives, and training data. The practical implication is sharper: decide in advance where the agent must hesitate.

Map one real customer task. Find the moments where two sensible people could reasonably want different things. Give each moment a rule: proceed, show, or ask. Then test the wording with customers. The goal is not to make every action ceremonial. It is to keep the person’s purposes from becoming background furniture.

Use The Courtesy Map before an agent becomes the most confident member of the service desk.

Continue with Write the Question Before the Claim and Every Brand Has a God Test.