Actions vs. triggers: what's the difference? Guide
Actions are things your assistant and agents can do inside a connected app. Triggers are events that start an agent on their own.
Last updated July 16, 2026
On this page
Every integration you connect brings two kinds of capabilities: actions and triggers. Understanding the difference helps you decide what your assistant is allowed to do and when your agents should run on their own.
Actions: things that get done
An action is a single operation the assistant or one of your agents can perform inside a connected app. Sending an email, creating a calendar event, adding a contact to your CRM, drafting an invoice, or looking up an order are all actions.
Actions are how work gets carried out. When you ask the assistant to do something in chat, or when an agent completes a task, it does so by calling actions on the apps you have connected.
You stay in control of which actions are available. When you connect an app, you choose exactly which of its actions the assistant is allowed to take, shown as a checklist you can select all of or narrow to just the ones you trust. You can revisit that list any time to add or remove enabled actions.
Tip: Connecting a powerful app doesn't mean granting everything. If you only want the assistant to read from an app, enable just its read actions and leave the rest turned off.
Triggers: events that start an agent
A trigger is an event in a connected app that can automatically start one of your agents. Instead of you asking for something, the app tells Praxivara that something happened, and an agent begins its work in response.
Common examples include a new e-commerce order, a change to a record in your CRM, a bank event, or an inbound text or call. When one of these events occurs, the agent you've set up runs without anyone needing to prompt it.
Triggers are what let an agent work on its own. If you want a workflow to happen automatically rather than on request, you need a trigger to start it.
The key difference
| Actions | Triggers | |
|---|---|---|
| What it is | An operation the assistant or agent performs | An event that starts an agent |
| Direction | Praxivara does something in the app | The app notifies Praxivara that something happened |
| Who starts it | You, in chat, or an agent running a task | The event itself, automatically |
| Example | Create a calendar event | A new order arrives and an agent begins |
Not every app has triggers
Every integration offers actions, but only some offer triggers. Many apps are designed purely for the assistant and your agents to act inside, with no automatic events to start a workflow.
To see what a specific app supports, open its capabilities view from the Integrations page. There you can browse the full list of actions and check whether the app also exposes any triggers. You can search within the list to find a specific operation and gauge whether the app fits a workflow you have in mind.
How to review an app's capabilities
- Go to your Integrations page and find the app you're interested in.
- Open the app to view its capabilities.
- Browse the Actions list to see everything the assistant and your agents can do inside it.
- Look for a Triggers list to see whether the app can start an agent automatically. If there isn't one, the app offers actions only.
- Use the search box to jump to a specific operation you have in mind.
Putting it together
Think of actions as the hands that do the work and triggers as the signal that tells an agent to begin. A well-built agent often uses both: a trigger starts it when something happens, and actions let it respond, whether that means updating a record, sending a message, or completing a task on your behalf.
If the assistant ever tells you an action isn't enabled, you can turn it on from the app's action settings. And if you want a workflow to run without being asked, look for a trigger on the relevant app to set it in motion.