Cortex XSIAM provides a comprehensive overview of the AI assets within an organization, designed to ensure AI security by offering tools to review and prioritize AI risks effectively.
**AI assets inventory**
You can view all AI assets in your environment, regardless of deployment mode or cloud provider. Navigate to **Inventory** → **All Assets** → **AI** to access the inventory. The top of the page features interactive widgets that summarize your AI security posture:
* **Assets at risk:** The number of AI assets with detected vulnerabilities or misconfigurations.
* **Cloud provider breakdown:** A summary of AI resources distributed across your cloud environments (AWS, Azure, GCP).
* **Sensitive assets:** The number of AI assets handling classified or sensitive data.
* **Sensitive assets discovered last 7 days:** A trend metric for newly exposed sensitive assets.
**AI asset categories**
The AI inventory is organized into the following categories to help you quickly filter your resources
* **All AI Assets:** An aggregated view of all AI resources.
* **Agents** AI agents (such as AWS Bedrock Agents) that use models and tools to perform tasks and assist primary models.
* **Datasets:** Collections of data used by AI models. This includes Training datasets used to teach the model how to process information, and Inference datasets used to provide real-world data during the model's inference phase such as for Retrieval-Augmented Generation, or RAG.
* **Models:** The trained machine learning models themselves , whether managed by a cloud provider or self-managed on your own cloud infrastructure.
* **Model Endpoints:** The interface through which applications interact with the AI model, acting as an access point for sending inputs and receiving generated outputs.
* **Software packages** The underlying AI-related software packages and SDKs used by developers to build these systems.
**Expanded AI asset information**
Clicking an AI asset in the inventory table opens a detailed asset card with specialized tabs for deep inspection:
* **Overview:** Summarizes highlights, properties, and asset identifiers.
* **Identity:** Provides an aggregated view of the permissions associated with the asset, mapping the relationships and displaying the identities that can access it.
* **Data:** Displays data profiles such as CCPA, Financial, and PCI, as well as data patterns found within datasets related to the asset.
* **AI Ecosystem:** Visualizes the "Asset Story," providing a topological map of how the AI components interact. For example, mapping the foundational Model connected to the Model Endpoint, tied to the Inference Dataset.