Artificial Intelligence in eSpatial’s Territory Manager

The Territory Manager application in eSpatial employs a structured AI architecture designed to ensure secure, controlled, and auditable interactions between client applications and AI services.

• Internal Security Perimeter: All message routing, workflow orchestration, and tool execution occur within eSpatial's internal environment. Intermediate data processing remains fully contained within our security perimeter.
• Controlled Access Gateway: Client interactions enter through a dedicated proxy interface that enforces security policies and acts as a strict barrier between user environments and AI services.
• Prohibited External Integrations: To maintain tight infrastructure boundaries, third-party or external client connections to internal AI execution tools are currently blocked.

• Stateless Response Generation: Natural language processing is handled externally via Anthropic through a commercial agreement. Customer inputs are processed transiently to generate immediate responses and are deleted immediately after completion.
• Zero Model Training: Customer data is contractually protected and never stored, retained, or utilized by external providers to train, fine-tune, or improve public or third-party AI models. eSpatial does not use customer data to train AI models.

• Disabled by Default: AI features are turned off by default across all customer organizations and require explicit administrative activation.
• Logical Tenant Isolation: Customer data is strictly segregated within a multi-tenant architecture. If an organization does not activate AI features, no data from that organization is ever transmitted to external AI services.
• Fine-Grained Permissions: Administrators maintain complete control over which individual users or roles have access to AI capabilities.

• Short-lived Authentication: Interactions generate temporary, time-bound tokens to validate actions internally within the appropriate security context.
• Token Isolation: External language models never receive, view, or interact with user authentication credentials or security tokens at any stage.

This multi-layered approach guarantees that organizations retain full governance over AI adoption while ensuring sensitive operational and spatial data remains under eSpatial's direct control.

• Self-hosted Monitoring: Activity traces and quality metrics are logged entirely within eSpatial's internal network for operational support and system troubleshooting.
• Strict 90-day Retention: Log visibility is restricted to authorized personnel via role-based access, and all operational trace logs are automatically deleted after a maximum of 90 days.