Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.tess.im/llms.txt

Use this file to discover all available pages before exploring further.

The Google Calendar step connects your agents directly to Google Calendar, allowing them to list events within a time range or automatically create new events. With it, Tess performs real actions in your calendar without manual intervention.

​
What is this Step?

This integration provides two actions within AI Studio:
  • Get Events: Reads events from a specific time range in the calendar and brings this information into the agent’s context.
  • Create Event: Creates a new event in the calendar with full details: title, description, date, time, guests, and notifications.
Image

​
How Steps work in Tess: read before configuring

This is the most important point to use the integration correctly: All App Integration steps run before any interaction with the user. This means that when a user opens the chat and starts talking to the agent, the steps have already been executed. Therefore, the agent cannot:
  • Wait for user approval to create the event
  • Use information provided mid-conversation to trigger the step
What the agent can do:
  1. Use User Inputs that were filled before the conversation started
  2. Use pre-configured fixed information in the step
  3. Use User Decision, which also collects the required data before the conversation

​
Practical summary:

ScenarioWorks
Step creates event with User Inputs data (filled before the chat)YES
Step creates event with fixed URL or configured dataYES
Step waits for user approval in chat to actNO
Step uses a response given mid-conversation to create an eventNO

​
Where to find it

  1. Go to AI Studio
  2. Click Add AI Step
  3. In Select Step Category, choose App Integration
  4. In Choose an App, select Google Calendar
  5. In Select Step Type, choose Create Event or Get Events

​
How to use (Quickstart)

Configure the following fields:
FieldWhat to fill in
Step NameInternal name of the step (e.g., Create onboarding meeting)
Event TitleEvent title (e.g., variable {{nome_cliente}})
Event DescriptionMeeting agenda or description
Start Date / Start TimeStart date and time (DD/MM/YYYY and HH:MM)
End Date / End TimeEnd date and time
TimezoneTime zone (e.g., UTC-03:00 for Brasília)
Send NotificationsYes to send email invites to participants
Calendar Owner EmailEmail of the Google account that will receive the event
Guest EmailsGuest emails separated by commas
NOTE:All fields marked with User Decision can be dynamically filled via inputs or fixed in the step configuration.
Image
Configure the following fields:
FieldWhat to fill in
Step NameInternal name of the step (e.g., Fetch weekly schedule)
Start DateStart date of the period to query
End DateEnd date of the period
TimezoneCalendar time zone
Calendar Owner EmailEmail of the calendar to be queried
The events found in this period will be available in the agent’s context to be used in the response or in other steps.
Image

​
Deeper explanation

Since steps run before the conversation, the usage logic changes. The agent does not create events “mid-conversation” — it starts the conversation with the event already created (or with the data already collected). The correct design is: 
User Inputs (collected beforehand)
        ↓
   Step is executed
(Create Event or Get Events)
        ↓
   Agent starts the conversation
 already with the event created or
  with the calendar data
Combining Get Events + Create Event:\ If you want an agent that “checks the calendar and schedules the best time,” ideally both steps should run in sequence with predefined data — for example, fetching the current week’s events and then creating an event in the first available slot. This type of logic works well in Text Agents triggered programmatically or via integration with another platform (e.g., Zapier, Make, or N8N).

​
Practical examples

Onboarding agent that automatically schedules a meeting

  • The user fills in the data at the beginning (User Inputs): name, email, available date and time.
  • The Create Event step uses these variables and creates the event before the chat starts.
  • The agent opens the conversation already confirming: “Hello! Your onboarding meeting has been scheduled for [date] at [time]. Here’s what we’ll cover…”_

Daily schedule briefing

  • A Text Agent is automatically triggered (via Make, for example) every morning.
  • The Get Events step fetches all events of the day.
  • The agent generates a structured summary of the schedule and sends it via email or Slack to the user.

Trigger via external form

  • Integration via N8N or Zapier triggers the Tess Text Agent with form data as inputs.
  • The Create Event step creates the event and sends the invite automatically.
  • No human interaction is required in the process.
Best practices
  • User Inputs are your allies: Whenever event data is variable (name, date, email), create a corresponding User Input and map it in the step.
  • User Decision is for simple cases: Use it when the user needs to provide the data before the agent starts — ideal for Text Agents.
  • Always define the Timezone: Events without the correct time zone may be created at the wrong time, especially if the calendar is in another state or country.
  • Separate emails correctly in Guest Emails: No spaces after commas. Example: {{email_cliente}},name@company.com
  • Name your step clearly: With multiple steps in the agent, clear names prevent confusion during maintenance.

​
Pay attention to the following points:

  • Mandatory authentication: Google Calendar must be integrated with your Tess account before using the step.
  • The step always runs: Regardless of the conversation context, the step will execute when the agent starts. Ensure input data is always available.
  • Credit consumption: Each step execution consumes credits from your plan.
Google Calendar turns Tess into an intelligent scheduling engine — but the key is understanding when it acts. With the right data available before the conversation begins, your agent creates events, reads schedules, and confirms meetings without any manual intervention.