​
Introduction

Humans represent your users that have a relationship Mindsets SDK or API. Humans are grouped into accounts automatically inside the Agent Management Portal, based on your existing permissioning system.

​
What information is stored about humans?

Mindsets founding team came from starting and scaling the largest UK & US based legal technology SaaS company, working with 90% of the fortune 1000. As such, the most advanced and modern security principles have been built from the ground up at Mindset to ensure human data is secure. All authentication (and storage of authentication tokens and user’s personally identifiable information) is also handled by Google. Furthermore, the authentication tokens are used in the database security rules; thus, the databases are secured at the user level. This results in the highest data security achievable in modern cloud computing environments.

​
Agent Memory Data Collection

When Agent Memory is enabled in your application, agents can collect contextual information about humans during conversations to enhance future interactions. This data collection operates under strict compliance frameworks and provides multiple layers of control:

​
Types of Information Collected

When enabled, agents may collect contextual facts, including:
  • Personal Preferences: Communication style, working hours, preferred contact methods
  • Professional Information: Job roles, responsibilities, expertise areas, and team relationships
  • Goals and Objectives: Project priorities, learning preferences, professional goals
  • Communication Patterns: Interaction history, preferred information formats
  • Organizational Context: Team structures, reporting relationships, departmental information

​
Data Storage and Security

  • Structured Storage: Facts are stored in Firebase subcollections with complete client segregation
  • Audit Trails: Every fact collection, modification, and deletion is logged with immutable timestamps
  • Classification System: Facts are categorized (Identity, Professional, Personal, Goals, Preferences, Communication)
  • Context Preservation: Original conversation context is stored with each fact for transparency
  • Agent Attribution: Every fact includes information about which agent collected it and when

​
Compliance Framework Integration

Agent Memory fully complies with major international data protection frameworks: GDPR Compliance (Articles 15, 16, 17, 20, 22):
  • Right to Access: Complete visibility of all collected facts through admin interfaces
  • Right to Rectification: Direct editing capabilities with full audit trails
  • Right to Erasure: Individual and bulk fact deletion with verification processes
  • Right to Data Portability: One-click export in machine-readable formats (CSV/JSON)
  • Automated Decision-Making Controls: Human oversight for all agent recommendations
EU AI Act Compliance (Articles 12, 18, 19):
  • Automatic Logging: All agent interactions logged with 6+ month retention
  • Transparency Requirements: Complete documentation of AI decision-making processes
  • Human Oversight: Meaningful intervention capabilities for all automated processes
CCPA/CPRA and US State Privacy Laws:
  • 45-day Response Times: Automated processing for all consumer rights requests
  • Data Correction: Individual fact modification with comprehensive validation
  • Complete Deletion: Bulk erasure capabilities with third-party notification

​
Human Control and Interaction Options

Direct Agent Interaction: Humans can make natural language requests directly to agents:
  • “Show me what information you have stored about me” (Right to Access)
  • “Please update my job title to Software Engineer” (Right to Rectification)
  • “Delete all my location preferences” (Right to Erasure)
  • “Send me all my data including what agents learned about me” (Right to Data Portability)
Administrative Interface Access:
  • Professional Management: Administrators can access comprehensive fact management through Human section → Agent Memory tab
  • Bulk Operations: Export all facts, delete all facts, or manage individual entries
  • Audit Trail Access: Complete modification history with timestamps and admin attribution
  • Compliance Reporting: Built-in tools for regulatory audit support and response time tracking
Opt-Out Controls:
  • Individual Control: Each human has a dedicated “Allow Fact Collection” setting
  • Granular Permissions: Humans can opt out at any time while preserving existing data or requesting complete deletion
  • Transparent Processing: Clear information about what data is collected and how it’s used

​
Where are humans found in Mindset?

To view and configure the Humans in your app, simply click on the “Humans” tab on the left-hand navigation bar. Here you will be able to see a list of all humans. Note: Humans cannot be created by an administrator inside of an application, end-users must create their own personal accounts and verify using their email address to access the Agent Management Portal or they must be created via Mindset APIs.