Enterprise AI discovery systematically identifies, catalogs(meta-data) AI agent related assets like models, tools and prompts across an organization’s AI landscape.
Enterprises are increasingly adopting AI solutions across various platforms, including AWS Bedrock, Google Cloud Platform (GCP), and Microsoft Azure and other SaaS platforms like Databricks, Salesforce etc. AI assets are scattered across multiple hyperscalers, leading to a lack of centralized visibility and control.
The administrator can configure discovery integrations with hyperscaler and SaaS environments to offer a unified overview of enterprise AI assets and their connections. This enables the AI steward to monitor AI systems such as AI Agents being created and deployed within the enterprise ecosystem, as well as track associated assets such as models, prompts, and tool details.
- Enables discovery of AI assets via integrations with: AWS Bedrock, SageMaker, Agent Core, Azure Foundry, Copilot Studio, GCP Vertex AI, Salesforce, n8n, Databricks, Snowflake, Hugging Face, LangGraph etc.
- Centralizes agents, models, prompts, and tools into one registry
- Aggregates assets across AWS, Azure, GCP, ServiceNow, and SaaS platforms
- Provides a single consolidated view of AI assets and their relationships
- Uses Service Graph Connectors framework for scalable integrations
Fixes
- Corrected asset owner assignment in domain-separated instances
In instances using domain separation, the automation that assigns an asset's "Managed by" field from the asset owner's email could, in rare cases, match a user in the wrong domain if the same email address existed on active users in more than one domain. The lookup is now scoped to the asset's own domain, ensuring the correct user is assigned.
Features
- Automatic "Managed by" assignment from asset owner email
Assets can now have their "Managed by" field automatically populated based on the asset owner's email address, when that email is captured as an additional attribute on the asset. The system resolves the email to an active user and applies it to the corresponding asset record. Includes safeguards to skip invalid or unmatched data so existing records are never overwritten incorrectly.
Maintenance
- General release preparation and dependency/version updates for the September 2026 release cycle.
Allow for existing AI Native SKU customers
With AICT app for AICT customers
The AI Discovery plugin (sn_ai_disc) has the following dependencies:
Plugin Dependencies:
- com.glide.hub.action_type.datastream
App Dependencies:
- sn_ai_governance
- sn_sgc_central