PlatformConnectivityFlowsLearn

AI Agent & AI Action in Flows

Power your flows with intelligence; choose between full agent logic or quick single-step actions.

When you log in through live.fastn.ai into your Fastn account, navigate to Flows and create a new flow by clicking on Create Flow.

Building your AI Agent On Chat Request

  • Select the starting point for your flow. For example, if you want to build your AI agent in Fastn through chat message, select On Chat message, name your flow, and click on Build to start.

  • Click on the "+" symbol within your Flow canvas.

  • In the component search bar, type “AI Agent” and select it.

  • In the AI model section, select an AI model of your choice.

  • Next, from the model dropdown, choose the AI model best suited for your use case (e.g., Gemini 2.0).

  • Under the Prompt section, define the input the agent should use, for example: {{input.chatInput}}

  • Adjust your Session ID, Memory Table Name, and Memory Content Limit as required.

  • Click Save to confirm your configuration.

The Fastn AI Agent is in the beta category and under trial.

  • You will now need to get into the loop of your connected AI Agent to select your apps to connect and set up necessary actions.

  • Select your desired app by searching for it in the Connector and connecting your account.

Example Use Case: Fetching your last Google Doc with one page and sending it to a Slack Channel

  • Select the Google Docs Connector and endpoint getDocs in this case, and click Next. In the next Connect section, connect your Google account to Fastn. Configure your action in the next step, and you will then see the getDocs tool connected in your Flow.

  • Similarly, connect your Slack app with the sendMessage endpoint. In the next Connect section, connect your Slack account to Fastn. Configure your action in the next step, and you will then see the sendMessage tool connected in your Flow.

Inside each connected tool, you can add or delete individual parameters and map them to input values that are static or defined automatically by the model, as shown below. You can learn more about data mapping from Data Mapping in Flows

AI Action

AI Action is your tool in Fastn to trigger action in your connected apps based on a single prompt without any model selection or complex flows.

Building your AI Action on API Request

  • Select the starting point for your flow. For example, if you want to trigger any task in Fastn through an API request for seamless system integration, select On API Request, name your flow, and click on Build to start.

  • Inside the flow search and add the AI Action flow component.

  • In the AI Action connector, click on Add Actions to select relevant apps for connection and to update the necessary actions [functionalities] for your task.

Example Use Case: Sending a Message to your Slack Channel

Search for the Slack Connector and enable the Send Message Action after authenticating your Slack account.

Your connector will appear as Connected after successful authentication.

You can now write a simple prompt with the message you want to send with the Channel Name e.g,

Send a message test to Slack Channel ai_testing

Click Save and test run your flow by clicking Test from the top-right corner.

You can deploy the flow to any of your selected environments to go LIVE.

You will now see your flow ran successfully and the message would be sent to the respective Slack channel.

PreviousData Mapping in FlowsNextUsing Templates

Last updated

Was this helpful?