How an app migrates to an agent
When migrating from Apps to Agents, each App component maps directly to its Agent equivalent to ensure seamless functionality.
Building blocks of an AI application
-
Name
- App Display Name → Agent name
- Direct one-to-one mapping with no changes
-
Icon
- App Icon → Agent Icon
- Visual representation remains identical
-
Description
- App Description → Agent Description
-
Instructions
- App Instructions → Instructions for the “Plan & Execute” step in the Agent (see screenshot below)
-
Tools
- App Tools → Tools for the “Plan & Execute” step in the Agent
- All configured tools and their customization are preserved as-is.
Additional migration details
- Permissions and Access:
- App library audience (Users, Departments) → Agent VIEWER roles
- App Editors → Agent EDITOR roles
- The creator is initially assigned as the Agent Owner. Creator history remains unchanged if ownership later transfers.
- Configuration:
- Prompts → Converted to conversation starters for the agent.
- Integration:
- Slack Settings are preserved as-is and can be accessed from the “Share menu” in the Agent builder.
ℹ️ The migrated agent will refer to itself asAgent: <name>in Slack in its responses instead ofApp: <name>.
- Slack Settings are preserved as-is and can be accessed from the “Share menu” in the Agent builder.
Example: Sample Agent (migrated from an App)
