AI PROPOSES
An AI proposes an action.
DECISIONAUTHORITYJOIN EARLY ACCESSDECISION AUTHORITY PERSONAL GUARDIAN
Your AI can help you.
It should not control your authority.
Personal Guardian creates a boundary between AI recommendations and actions taken in your name.
AI proposes. You retain authority.
ILLUSTRATIVE EXPERIENCE
AI PROPOSAL“Send this message.”
PERSONAL AUTHORITY FIREWALLPERSONAL GUARDIAN
AUTHORITY CHECKYOUR RULES · YOUR AUTHORITY
DECISIONASK ME
ASK METhe proposed action pauses pending your authorization.
THE AI DOES NOT GRANT ITSELF PERMISSION.
THE SHIFT / 01
AI ASSISTANTS ARE
BECOMING AGENTS.
AI systems are moving from answering questions toward taking actions. Greater capability makes one separation essential.
INFERENCEWHAT THE AI THINKS YOU WANT
≠AUTHORITYWHAT YOU HAVE ACTUALLY AUTHORIZED
PERSONAL AUTHORITY FIREWALL / 02
A CONTROL LAYER BETWEEN
INTENT AND ACTION.
Your chosen rules govern whether a proposed action can proceed, must wait for approval or must stop.
ACTION IDENTIFIED
Personal Guardian identifies the requested action.
RULES EVALUATED
Your rules and permissions are evaluated.
APPROVAL REQUESTED
Sensitive actions can require your approval.
UNAUTHORIZED ACTIONS STOP
Actions without authority are blocked.
AUTHORIZED ACTION CONTINUES
Only an approved path may continue.
ILLUSTRATIVE EXPERIENCE / 03
CAPABILITY CAN BE HELPFUL.
AUTHORITY MUST BE EXPLICIT.
These scenarios illustrate the intended authority model. They do not represent confirmed integrations or general availability.
AI PROPOSES
“Send this message.”
GUARDIAN
PAUSED FOR YOUASK ME
AI PROPOSES
“Complete this purchase.”
GUARDIAN
PAUSED FOR YOUASK ME
AI PROPOSES
“Move tomorrow’s meeting.”
GUARDIAN
MAY PROCEEDALLOW
AI PROPOSES
“Share these account details with another service.”
GUARDIAN
TERMINATEDBLOCK
AI PROPOSES
“Change this account setting.”
GUARDIAN
PAUSED FOR YOUASK ME
ONE PERSON, MANY AIS / 04
YOUR AUTHORITY SHOULD FOLLOW YOU
ACROSS AI SYSTEMS.
The Personal Guardian layer remains constant as different AI assistants propose actions.
AI 01AI 02AI 03AI 04
PERSONAL GUARDIANYOUR RULES · YOUR APPROVALS · YOUR AUTHORITY
TWO PRODUCTS / ONE PRINCIPLE / 05
Organizational authority.
- FOR
- Organizations and regulated systems
- AUTHORITY SOURCE
- Organizational policy + authorized humans
- BOUNDARY
- AI proposal → governed enterprise execution
Individual authority.
- FOR
- Individuals
- AUTHORITY SOURCE
- The individual
- BOUNDARY
- AI proposal → personal permission → external action
PRODUCT PHILOSOPHY / 06
AUTHORITY SHOULD
BE EXPLICIT.
Personal Guardian is designed around making consequential AI actions visible, bounded and subject to the individual’s chosen rules.
- Actions remain distinct from recommendations
- Sensitive actions can require direct approval
- Actions without authority stop before execution
PERSONAL GUARDIAN EARLY ACCESS
KEEP THE INTELLIGENCE.
KEEP THE AUTHORITY.
Personal Guardian is in development. Join the list for product updates without any claim of current availability or a launch date.
FOR ORGANIZATIONS — EXPLORE DECISION AUTHORITY ENTERPRISE