# How to use the AI Agent action to automate tasks in workflows > Learn how to use the powerful AI Agent action in Futuria CRM workflows to automate complex tasks with natural language instructions. Author: Futuria Support Team · Published: June 3, 2026 · Reading time: 9 min read Source: https://futuriamarketing.com/en/supporto/guida-azione-agente-ai-workflow/ --- ## What is the AI Agent action in workflows? The **AI Agent** action in workflows helps you automate complex tasks using natural language instructions and a selection of tools. Instead of building every single step manually, you can configure a single AI Agent action that decides what to do, which tools to use and which inputs to supply at runtime. For example, instead of creating separate workflow actions to find a contact, generate a booking link, format a date and send follow-up messages, you can configure a single AI Agent action to handle the whole process. ***Note:** The AI Agent is a **premium** workflow action and may require an add-on before you can use it.* ## Key benefits of the AI Agent action - **Autonomous execution:** the agent decides which tools to use, in what order and with which inputs, based on your instructions. - **Faster workflow building:** a single AI Agent action can replace several manually configured steps, for the use cases it supports. - **Flexible configuration:** you can start from scratch or use a predefined template and customise the configuration. - **Support for subsequent steps:** the agent can return output in text or JSON format for use in later stages of the workflow. - **Built-in visibility:** the execution logs let you review the agent's reasoning steps, tool usage and token consumption. ## AI Agent action details These settings control the AI Agent's behaviour and provide the essential inputs for configuring the action correctly: - **Action name:** add a custom name for the action. The default name is *AI Agent*, but renaming it to match the use case makes the workflow easier to manage. - **Template:** choose a predefined template to pre-fill the **Instructions** and **Tools** fields, or select **Build your own** to start from scratch. Some templates require a specific workflow trigger. - **Instructions:** define what the agent should do, how it should make decisions and which rules it should follow. This is the most important part of the configuration. - **Improve prompt:** improve roughly written instructions by restructuring them into a more detailed, organised prompt. You can undo the improvement if you need to. - **AI model:** select the artificial intelligence model that powers the agent. - **Tools:** choose the actions the agent is allowed to perform. - **Conversation memory:** use a running summary of past executions for the same contact at this step, so that the agent has context. - **Output format:** choose how the agent should format its response. ![AI Agent action settings in a Futuria CRM workflow](https://jumpshare.com/share/szPJPYdoXAC0h1H9JM2L+/GIF+Recording+2026-04-06+at+19.25.41.gif) ## Templates Templates help you get started faster by pre-filling common instructions and tools for specific workflow scenarios. They are useful when you want a starting point that matches a known use case, which you can then customise for your business. Templates are predefined agent configurations that populate the **Instructions** and **Tools** fields. If a template requires a specific trigger and that trigger is not already present, the template will appear unavailable until the required trigger has been added. **The available templates include:** - Build your own - Form lead follow-up - Missed appointment recovery - Facebook lead nurturing - Stalled deal chaser - Lead research and enrichment - New appointment: enrich and confirm - Quiz lead scoring and routing - Call transcript summary and actions - Task creation from won deals - Instagram comment to DM - Facebook comment to DM - Lead pipeline tracking ![List of the templates available for the AI Agent action](https://jumpshare.com/share/LHazacjOLKM6HDeJLKVi+/Screenshot+2026-04-06+at+20.07.16.png) ## Writing the instructions Well-defined instructions help the agent make better decisions, use tools correctly and produce more reliable results. The clearer your guidance, the easier it will be for the agent to carry out the task as intended. The **Instructions** field defines what the agent does, how it reasons and how it uses the tools you provide. You can keep the instructions simple or make them more detailed depending on how complex the task is. Good instructions usually include: - A statement of the role - Context about the data available from the trigger - A step-by-step workflow - Decision logic with if/then rules - Messaging guidelines - Rules or constraints on what the agent should avoid doing If you are not sure how to structure the prompt, enter your instructions in natural language and use **Improve prompt** to reorganise them. The editor also supports rich text formatting, including headings, bold, bulleted lists, numbered lists, code blocks and quotations. ![Text editor for writing the instructions for the AI Agent](https://jumpshare.com/share/4UnOEIvYfUvZh3nxJgH4+/GIF+Recording+2026-04-07+at+18.39.45.gif) ## Configuring the tools Tools determine what the agent is allowed to do at runtime. Giving the agent the right tools and deciding which fields it can control helps you balance flexibility with predictability. Click **+ Add tool** to open the selector. Tools are organised by category and you can also search for them by name. Each tool you select appears in the configuration panel and can be configured individually. When you configure a tool, you can choose how the field values are populated: - **Let AI decide all field values: ON**. The agent generates the values for all fields at runtime, based on your instructions and the workflow data. - **Let AI decide all field values: OFF**. You can set some values manually, allowing the agent to fill in the remaining empty fields. Each field also has its own toggle, so you can switch individual fields between AI-generated values and manually configured values. ![Configuring the tools for the AI Agent with manual and automatic options](https://jumpshare.com/share/Vuj5V1dr8ZpGjWxqqyfg+/GIF+Recording+2026-04-06+at+20.17.36.gif) ## How to set up the AI Agent action in a workflow A structured set-up makes it easier to choose the right template, write clear instructions and give the agent the tools it needs to complete the task successfully. 1. Go to **Automations > Workflows**. ![Navigating to the Workflows section in Futuria CRM](https://jumpshare.com/share/0JWEsWQU9NawnWRXArkh+/Screen+Shot+2026-04-06+at+21.20.01.png) 2. **Create** a new **workflow** or **edit** an **existing** one. ![Editing an existing workflow](https://jumpshare.com/share/BXEsCvFYK7i5WsZpbUNj+/Screen+Shot+2026-04-06+at+21.21.45.png) 3. Add a relevant trigger such as **Form submitted**, **Contact created** and so on. ![Selecting a trigger for the workflow, such as 'Contact Created'](https://jumpshare.com/share/7cvd7EvdDe5EzEzze4Zm+/GIF+Recording+2026-04-06+at+21.28.31.gif) 4. Click **+ Add** (or select the **Select action** placeholder) in the workflow canvas. 5. Select **AI Agent** from the **AI Workflows** section. 6. Enter an **Action name**. 7. You can select **Build your own** to configure the action from scratch. 8. Add your **Instructions**. 9. Click **Improve prompt** to improve the structure of the prompt. 10. Choose the **AI model** you want to use. 11. Click **+ Add tool** and select the tools the agent can use. 12. Configure each tool, specifying whether the AI should decide the field values or whether certain values should be set manually. ![Adding and configuring the tools for the AI Agent action](https://jumpshare.com/share/rL6dUAKMPEV4IVjmBCOG+/GIF+Recording+2026-04-07+at+19.45.24.gif) 13. Alternatively, you can select a **Template**. Templates are predefined configurations that populate the **Instructions** and **Tools** fields. If a template requires a specific trigger that is not present, it will appear unavailable until that trigger has been added. ![Selecting a predefined template for the AI Agent](https://jumpshare.com/share/K0Y1edlbfPxhdnRFeVYS+/GIF+Recording+2026-04-07+at+19.51.17.gif) 14. Open the **Advanced options** (optional) to enable **Conversation memory** or configure the **Output format**: - **Conversation memory:** helps the agent maintain continuity across repeated executions for the same contact at that step. It is useful for recurring workflows in which the same contact may re-enter the step and the agent should adapt based on previous interactions. - **Off by default:** when enabled, the agent uses a running summary of past executions for the same contact at that step. - Use it for **recurring** or **multi-contact** workflows (for example periodic follow-ups). For one-off workflows (for example a single form follow-up), it can usually stay switched off. - **Output format:** controls whether the agent returns data that later workflow steps can use. Choosing the right output type is useful when the workflow needs to reference the agent's result after it has run. - **None:** performs the actions but does not return structured output for subsequent steps. - **Text:** returns a free-text response. When selected, an **Output description** field appears so you can describe what the response should contain. - **JSON:** returns structured data based on a defined schema. You can add properties with a name, a type and an optional description to make the output easier to reference later in the workflow. 1. **Save** the action. ![Saving the configured AI Agent action](https://jumpshare.com/share/GzSHLAMZcSAIh5XoNLp7+/Screen+Shot+2026-04-07+at+19.53.43.png) ## Execution logs Execution logs make it easier to understand what the agent did during a workflow run. They are useful for validating behaviour, checking tool usage and troubleshooting unexpected results. After the workflow has run, open the **Execution logs** tab in the workflow builder and select an AI Agent execution to view its trace. The trace can include: 1. **Agent input:** the full input the agent received. 2. **LLM:** each model call used to reason about the next step. 3. **Tool executions:** every tool the agent called, including execution time and success or failure status. You can expand each step to review the input and output data. The log view supports both a table format and a raw JSON view. ![Viewing the execution logs for the AI Agent action](https://jumpshare.com/share/9QUweuvvPUpZZ9c21JUC+/_Umdeshfp6Wt-nH6PHIhYGP_Ws4HJVrE7w.png) ## Frequently asked questions ##### Is there a limit to the number of tools I can add? Yes. You can add up to **10 tools** per agent. Internal tools such as **Date calculator** and **Maths operations** do not count towards this limit. ##### Which AI model should I use? Use a general-purpose model for most use cases, choose higher "reasoning" levels for more complex logic, and lighter models for simple tasks where speed matters. ##### Can I use the AI Agent with any trigger? Yes. The AI Agent works with any workflow trigger. Templates may recommend specific triggers because their instructions are written for that context. ##### What does the "Improve prompt" feature do? Improve prompt restructures your instructions into a more organised, detailed prompt, including clearer steps and edge-case handling. You can undo the improvement if you need to. ##### When should I use the JSON output format? Use JSON when later workflow steps need to reference specific values from the agent's response, such as qualification results or other structured output. ##### How can I see what the agent did during a workflow run? Open the **Execution logs** tab in the workflow builder and select the AI Agent execution to review the trace, including model calls, tool executions, inputs, outputs and token usage. ##### Do templates lock me into a specific configuration? No. Templates pre-fill the **Instructions** and **Tools** fields, but you can edit the instructions, change the AI model, add or remove tools and update the advanced options after applying a template. ##### When should I enable Conversation memory? Enable it when the same contact may pass through the agent step more than once and the agent should adapt based on previous executions. For one-off workflows, it is usually not needed.