Action Tool

An AI agent that can only answer questions is useful. But an agent that can also create a work order, send a WhatsApp message to the assigned technician, and generate the shift PDF report — all from the same conversation — is transformative. Action tools are what allow the agent not just to respond, but to act.

What is it for?

Action tools connect the agent to the platform's operational functions. A technician reports a failure over WhatsApp, the agent logs it as a task, assigns it to the right department, notifies the on-call technician, and confirms to the user that everything is in motion — without anyone touching the dashboard keyboard.

Action tools allow the agent to:

• Create and assign maintenance tasks automatically
• Send WhatsApp messages to technicians or supervisors
• Send emails with or without file attachments
• Generate PDF reports with corporate branding
• Locate the nearest technician to a specific geographic location

How does it work?

Each action tool is linked to a system function. When the agent decides it needs to execute an action — because the user requested it, or because the situation requires it based on its prompt — it calls the tool with the necessary parameters. The function runs in the backend and the agent receives confirmation to include in its response.

The agent combines parameters from two sources:

• What the user said in the conversation (equipment name, problem description, priority)
• Default values configured in the tool (department, email account, report branding)

This allows the agent to act with complete context without having to ask the user for information that is already configured.

Available actions

How to use it?

Create an action tool

1. Go to Tools in the sidebar.
2. Click New Tool.
3. Select the type Internal action.
4. Choose the action from the list.
5. Configure the default values (see next section).
6. Click Save and assign the tool to the agent.

Configure the "Create and assign task" action

This is the most commonly used action in industrial maintenance agents:

The agent collects from the user the title, description, and urgency. Using the default values plus what it gathered from the user, it creates the complete task. The assigned technician receives the notification automatically.

Configure the "Generate PDF report" action

The report is produced with the corporate branding configured in the default values:

The agent builds the report content from the data it collected during the conversation — technician readings, observations, findings — and combines it with the preset branding to generate the PDF.

Configure the "Locate nearest personnel" action

This action finds the nearest technician to a geographic location in order of proximity: first in the same city, then in the same state, then in the same country. You can filter by department so it only searches technicians from the correct area.

This is especially useful for service companies with technicians distributed across multiple cities who need to assign work to the technician closest to the client.

Key benefits

• The agent acts without depending on a human opening the dashboard
• Default values eliminate unnecessary questions to the user
• Automatic notifications guarantee that the assigned technician is always informed
• PDF reports are generated with corporate branding without additional configuration per report
• A single conversation can chain multiple actions: create task + send WhatsApp + generate report
• The agent confirms to the user which actions were executed and with what result

Common use cases

Scenario 1: Automatic failure report The shift operator writes to the WhatsApp agent: "Compressor C-02 is vibrating heavily — I think it has a bearing problem." The agent creates a task "Bearing Inspection — Compressor C-02" with high priority in the Mechanical Maintenance department, sends a WhatsApp message to the on-call technician with the details, and responds to the operator: "Task #MAN-0924 created with high priority. Technician Miguel Torres has been notified and will confirm within 15 minutes." All in under 10 seconds.

Scenario 2: Shift report generated by the agent At the end of the shift, the supervisor tells the agent: "Generate the afternoon shift report." The agent queries completed tasks, alarms from the period, and equipment that triggered alerts. It generates a PDF with the company logo, shift activity tables, and sends it to the maintenance manager's email. The supervisor does not need to write the report manually.

Scenario 3: Smart assignment for a service company A maintenance service company receives a client request in Monterrey. The agent runs "locate nearest personnel" with city "Monterrey" and department "Industrial Maintenance." It finds 2 technicians in Monterrey and 1 in Saltillo. It creates the task assigned to the nearest technician, sends them a WhatsApp message with the client details, and confirms to the client the assignment with the technician's name and estimated arrival time.