Brainis

Autonomy

You choose how autonomous.

Nine questions, answered in order — the whole authority model, plainly.

The nine questions

1 of 9

What can Brainis see?

What you connect, scoped by sensitivity — and you can inspect any claim's sources.

Instrument

The full ladder.

L3 — Ask to act

May do

  • research prospectslowreversible
  • draft sequenceslowreversible
  • queue for reviewlowreversible

Must ask

  • send to new segmentsmediumirreversible
  • exceed $200/week budgetmediumirreversible

Never

  • contact existing customers' clientshighirreversible
  • change pricinghighirreversible

Anatomy

The Authority Contract.

Authority Contract

actor
who or what holds this contract
scope
the domain it applies to
verbs
the specific actions it grants
budget
AI capacity and external spend ceilings
risk ceiling
how much risk it may take unsupervised
reversibility
required, unless the gate says otherwise
window
how long it runs
verifier
who or what checks its work
expiry
when it lapses and must be renewed
UnderstandDecideDeliverVerifyLearn

Authority contracts govern every action agents take across the whole loop.

Hard question

What is L7, really?

Earned, scoped, revocable, expiring — autonomy that proved itself in the ledger, renewed like a license, never “the AI runs free.”
See the Trust Inspector →Back to the product map →