Task-specific permission
June is designed to request sensitive access because a task needs it, not because the assistant wants broad background visibility by default.
Privacy and trust
A personal assistant has to earn the right to know you.
June is designed around permissioned access, visible action, and user-controlled memory. The assistant becomes more useful with context, and that context stays more trustworthy when it is understandable, correctable, and removable.
At a glance
- Sensitive capabilities start with explicit user permission.
- Memory is built around review, correction, and deletion.
- Actions stay understandable before they affect the real world.
Permission model
June treats access as something the user grants, not something the product assumes.
A mobile assistant may need context from apps, messages, reminders, location, or the screen to be useful. June's trust model is to ask for access in context and make the purpose clear.
Visible follow-through
When June helps with a ride, order, message, or reminder, the user can understand what is happening and what will be submitted.
Memory with restraint
Memory is useful when it improves future tasks. June treats it as purposeful context, not a vague bucket of everything the assistant has ever seen.
Control and deletion
Users can correct saved context, remove it, and understand which details are shaping the assistant's behavior.
Product principles
A trustworthy assistant explains the boundary between help and control.
June's public privacy story is intentionally practical. People do not need abstract promises; they need to know what the assistant can access, why it needs that access, and how they can change their mind.
Clear purpose
Every sensitive capability connects to a user-visible purpose, such as completing a task, remembering a preference, or reducing repeated setup.
Understandable state
People do not have to guess why June suggested something. Important memory and context are easier to trust when they are explainable in plain language.
User approval
High-impact actions should keep the user in the loop. The stronger the action, the more important confirmation and visibility become.
No lock-in posture
A personal assistant works better when people can leave, delete, or reset context without turning memory into a trap.
Privacy FAQ
Questions people ask before trusting an AI assistant
Why does a personal AI assistant need permissions?
June can only help with phone tasks when the user grants the access needed for that task. The trust model is permissioned: sensitive access is requested clearly and remains under user control.
Can June's memory be changed or deleted?
June's memory is intended to be user-controlled. People can review important saved context, correct it, remove it, and decide what the assistant should not keep.
How does June make actions feel safe?
Actions are designed to be visible and understandable. When an assistant books, sends, orders, or changes something, users need enough context to know what is happening and stay in control of the outcome.