Apps that support AI agent imports

You can configure these AI agent provider types in Okta:

  • Embedded: AI agents that are built directly into the apps that your org already uses.
  • Standalone SaaS: Purpose-built, third-party AI products that operate as independent entities.
  • Homegrown (fully custom): Custom AI agents that are built in-house using raw code (like Python or LangChain).
  • Homegrown (agent builder platform): AI agents that are built using provider-specific agentic frameworks that handle the runtime, memory, and orchestration.
App integration Supported platform

Imported AI agent type

Resource connection support

Configuration guide

AWS Identity and Access Management Amazon Bedrock Agents Homegrown (agent builder platform)
AWS Identity and Access Management Amazon Bedrock AgentCore Homegrown (agent builder platform) Secure an Imported Amazon Bedrock AgentCore Agent.
Claude Claude Managed Agents Homegrown (fully custom and agent builder platform)
Datarobot Datarobot Agent Workforce Platform Homegrown (agent builder platform)
Glean Glean Agent Builder Standalone SaaS, Homegrown (agent builder platform)
Google Workspace Gemini Enterprise Agent Platform Homegrown (agent builder platform)
Microsoft Office 365 Microsoft AI Foundry Homegrown (agent builder platform)
Microsoft Office 365 Microsoft Copilot Studio Homegrown (agent builder platform)
Salesforce.com Salesforce Agentforce Embedded, Homegrown (agent builder platform)
ServiceNow UD ServiceNow AI Agent Studio Homegrown (agent builder platform)
Workday Workday Agent System of Record Embedded, Homegrown (agent builder platform)

To secure imported Homegrown: Agent builder platform AI agents that were built with custom code, complete these tasks:

  1. Configure imported AI agents
  2. Connect AI agents to resources
  3. Set up AI agent token exchange

To secure Homegrown: Fully custom or Homegrown: Agent builder platform AI agents that were built with custom code that don't have an available connector, complete these tasks:

  1. Add AI agents manually
  2. Connect AI agents to resources
  3. Set up AI agent token exchange