Skip to main content
An Org Admin sets up Auto once, for the whole organization. You choose which model answers at each tier, and Auto picks between them for each conversation.
Auto is included on the Business, Business Elite and Enterprise plans.

​
Choose your tier models and switch Auto on

1

Open the Models tab

Go to Org settings → Models and find the Auto card.
2

Choose the Everyday model

Everyday model answers routine messages, which is most of what your team sends. Parlé suggests the lowest-priced model your organization can use.
3

Choose the Standard model

Standard model answers demanding messages: long conversations, attached files, and tool use the Everyday model can’t handle.
4

Optionally, choose a Frontier model

Frontier model is your most capable model. Chat reaches it only when a person asks: with Deep, with Try with a stronger model, or in a project set to start at Frontier. Auto never chooses it automatically in chat. You can leave it empty.
5

Switch Auto on and save

Tick Switch Auto on and select Save Auto settings. Auto now appears at the top of the model picker for everyone in your organization.
Each tier needs a different model, and every model you choose must be allowed by your organization’s model policy. If a policy later blocks one of them, the card shows Blocked by your model policy and Auto falls back to the next tier down for affected conversations.

​
Make Auto the default

Out of the box, new conversations start on your organization’s default model, not on Auto. To make Auto the starting point:
  • For the whole organization: on the Auto card, select Make Auto the default. The model that used to be the default now shows a Default when Auto is off label. That is where new conversations go if Auto stops being available.
  • For yourself: anyone can choose Auto as their own default model in their profile.
If Auto is the default but can’t be used (for example, it has been switched off), new conversations open on the next default instead, with a one-line note saying why. Nothing is changed for you: switching Auto back on makes it the default again.

​
Set where Auto starts in a project

Some engagements always need a stronger model. A Project Admin or Org Admin can set a project’s Auto starting tier on the project’s Budget tab. Choosing a higher starting tier shows the price difference first, because it raises the cost of every Auto message in the project. Select Save starting tier to apply it. Budgets and the per-message spending limit still apply.
Frontier is offered only when your organization has a Frontier model. If the starting tier’s model is briefly unavailable, Auto uses the next model down and says so. The conversation moves back up at its next pause once the model can answer again.

​
If your plan changes

If your organization moves to a plan without Auto:
  • Auto stops choosing models. Each conversation on Auto stays on the model it last used, with a one-line note. Nothing is refused.
  • Your settings stay visible. You can still view your tier models and switch Auto off.
  • Nothing is deleted. Past answers keep their captions, and your usage reports still show how past spend was split by tier.

How Auto works

What members see, how Auto picks, and how to ask for a stronger answer.

Model policy

Decide which models can be used at all. Auto only picks inside these rules.