Action steps in flows
Action connectors enable flows to integrate with external systems and services, extending the capabilities of your Amazon Quick Flows beyond data retrieval and analysis. These connectors allow your flows to perform actions such as creating records, sending notifications, updating databases, and triggering workflows in connected applications.
Amazon Quick Suite Actions supports hundreds of actions across multiple third-party systems, enabling comprehensive automation and integration capabilities.
Prerequisites for adding action connectors
Administrators must configure action connectors in the Amazon Quick Suite console before they become available to Amazon Quick Flows creators. Only plugins and associated actions configured by administrators will be accessible for use in Flow creation. This includes third-party connectors, MCP connectors, and custom API connectors.
System requirements
Your Amazon Quick Flows environment must have the appropriate permissions and network connectivity to communicate with target systems. Verify that firewall rules, security groups, and network access control lists allow outbound connections to the services you plan to integrate. Additionally, confirm that your Amazon Quick Flows instance has sufficient resources to handle the additional processing load that action connectors may introduce.
Authentication credentials
Obtain the required authentication credentials for each service you plan to connect. This typically includes API keys, client IDs, client secrets, or other service-specific authentication tokens. Store these credentials securely using your organization's approved credential management system, ensuring they are accessible to your Amazon Quick Flows environment while maintaining security best practices.
Service account setup
Configure service accounts or application registrations in the target systems where your action connectors will operate. These accounts should have the minimum necessary permissions to perform the required actions while adhering to the principle of least privilege. Document the specific permissions granted to each service account for future reference and security audits.
Testing environment
Establish a testing environment that mirrors your production setup to validate action connector functionality before deployment. This environment should include access to test instances of your target systems, allowing you to verify connector behavior without affecting production data or processes.
Setting up Action steps in flows
To add action steps to your flow, follow these steps:
-
In the Flow builder, select the + Add step button.
-
From the menu, choose Application actions step.
-
In the Edit panel, choose your connector and your action.
-
Write your prompt for the action execution.
-
Save your configuration.
Using action connectors in flows
Action connectors integrate seamlessly into your flow design, appearing as actionable steps that can be configured and chained with other Flow components. The implementation process involves selecting the appropriate connector, configuring its parameters, and defining how it interacts with other elements in your flow. Amazon Quick Flows supports enhanced action flows that display eligible fields as dropdowns, reducing cognitive load and improving user experience.
Connector selection and configuration
Choose the appropriate action connector based on your integration requirements and the target system's capabilities. Each connector provides a configuration interface where you specify connection details, authentication information, and action-specific parameters. Flow creators can select from available plugins configured by administrators and choose specific actions from dropdown menus.
Data mapping and transformation
Configure how data flows between your Amazon Quick Flows and the connected systems. This includes mapping input parameters from your flow to the connector's expected format and defining how response data should be processed and passed to subsequent Flow steps. The mapping interface provides tools for data transformation, allowing you to modify data formats, apply filters, or perform calculations as needed.
Error handling and retry logic
Implement appropriate error handling mechanisms to manage potential connectivity issues, authentication failures, or service unavailability. Configure retry policies that define how the connector should respond to temporary failures, including retry intervals, maximum retry attempts, and escalation procedures for persistent issues.
Flow integration patterns
Design your flows to effectively utilize action connectors within the broader Flow logic. This includes determining the optimal placement of action steps, configuring conditional running based on previous step results, and implementing parallel processing where appropriate to optimize performance.
Authentication: 2 Legged OAuth vs 3 Legged OAuth
Action connectors support multiple authentication methods to accommodate different security requirements and integration scenarios. Understanding the differences between 2-legged and 3-legged OAuth helps you choose the appropriate authentication approach for your specific use case.
2 Legged OAuth
2-legged OAuth, also known as client credentials grant flow, provides server-to-server authentication without requiring user interaction. This method is ideal for automated processes where your flow needs to perform actions on behalf of the application rather than a specific user. The authentication process involves your application directly exchanging credentials with the target service to obtain an access token. This flow is suitable for service-to-service token authentication and is used when the identity of the end user does not matter, such as when accessing service-owned resources or for bot operations.
This approach offers simplified implementation and reliable automation since it doesn't depend on user presence or interaction. However, actions performed using 2-legged OAuth are typically associated with the application or service account rather than individual users, which may limit audit trails and personalization capabilities.
3 Legged OAuth
3-legged OAuth, also known as authorization code grant flow, involves the end user in the authentication process, requiring them to explicitly grant permission for your flow to access their account in the target system. This method is suitable when the identity of the end user matters, such as when accessing user-owned resources or when operations should be performed under the user's identity rather than a bot. This flow is supported by most ISVs who support OAuth and provides user-specific access with appropriate user context and permissions.
The 3-legged OAuth process requires users to authenticate with the target service and authorize your application's access during their first interaction with the flow. Subsequent runs can use stored refresh tokens to maintain access without repeated user intervention, provided the tokens remain valid and the user hasn't revoked access.
This authentication method provides better security and audit capabilities since actions are tied to specific user accounts. However, it requires more complex implementation and user interaction, which may not be suitable for fully automated processes.
Available action connectors
Amazon Quick Flows supports a comprehensive range of action connectors that enable integration with diverse systems and workflows. For the most current list of available connectors and their capabilities, see Working with integrations and Action integrations.
Connector capabilities and limitations
Each connector category offers different levels of functionality, support, and maintenance. Second-party connectors offer reliable integration with established partner services, while third-party connectors provide broad compatibility with varying levels of feature completeness.
When selecting connectors, consider factors such as authentication requirements, rate limiting, data format compatibility, and long-term maintenance commitments. Review the specific documentation for each connector to understand its capabilities, limitations, and best practices for implementation.