Skip to content
Karyam

Service Catalogs

Service Catalogs provide a curated and governed experience for consuming AI capabilities within Karyam.

Rather than exposing the full platform, Service Catalogs present only the AI services that have been explicitly published and approved for use.

This allows organizations to separate:

  • AI builders
  • AI operators
  • AI consumers

Karyam currently supports three primary user roles:

Role Purpose
Admin Manage the platform, governance, permissions, and infrastructure
Developer Build and maintain AI systems
User Consume published AI capabilities

A Service Catalog acts as the business-facing portal of Karyam.

It provides a simplified experience where users can:

  • Discover available AI services
  • Execute approved workflows
  • Interact with published agents
  • Track their own activity
  • Respond to approval requests

Users do not need to understand:

  • Models
  • Vector databases
  • Skills
  • MCP servers
  • AI infrastructure

They simply use the capabilities provided to them.


The Service Catalog surfaces only resources that have been explicitly published by administrators or developers.

These may include:

Business users can interact directly with approved AI agents.

Examples include:

  • IT Support Assistant
  • HR Assistant
  • Sales Copilot
  • Procurement Assistant

Users can execute workflows without understanding the underlying orchestration.

Examples include:

  • VPN Access Request
  • Employee Onboarding
  • Expense Approval
  • Purchase Request


Agents and AI Flows become available in the Service Catalog only after they have been published.

Publishing is intentionally simple and can be done directly from the resource list.

  1. Navigate to the Agents or AI Flows page.
  2. Locate the resource you want to publish.
  3. Enable the Publish toggle in the list view.

Once the toggle is enabled, the resource becomes available in the Service Catalog for users with the appropriate permissions.



Published services are organized into categories to improve discoverability.

Examples:

Human Resources
├── Employee Onboarding
├── Leave Assistant
└── Benefits Support
IT Services
├── VPN Access
├── Password Reset
└── Software Installation
Finance
├── Expense Approval
├── Budget Request
└── Purchase Orders

Categories allow organizations to present AI services in a familiar business structure.


Users can view their own activity directly from the Service Catalog.

This includes:

  • Previously executed runs
  • Workflow history
  • Agent interactions
  • Request status
  • Execution outcomes

Users only see activity that belongs to them.


If a user is responsible for approving actions, approval requests appear directly within the Service Catalog.

Examples include:

  • Manager approvals
  • IT approvals
  • Finance approvals
  • Security approvals

Approvers can:

  • Review requests
  • Approve actions
  • Reject actions
  • Provide feedback

Service Catalogs provide an important governance boundary.

Developers build systems.

Administrators govern systems.

Business users consume systems.

This separation allows organizations to scale AI adoption safely without exposing unnecessary complexity.


Agents
+
AI Flows
+
Categories
+
Approvals
↓
Service Catalog

The Service Catalog does not create AI capabilities.

It provides secure and governed access to AI capabilities that already exist within the platform.


Not every user needs access to AI infrastructure.

Most users simply need access to the right capability at the right time.

Service Catalogs transform AI systems into business services that can be safely consumed across the organization.