Introduction

Triggers define the conditions or events that activate a plugin within Agent Studio, enabling the Moveworks AI Assistant to respond intelligently to user needs or system events. By configuring triggers, you can ensure plugins are invoked in the right context—whether initiated by a user's natural language query, an external event, or a predefined schedule. This allows for seamless automation, improving efficiency and providing timely assistance across your enterprise systems.

Triggers bridge the gap between user interactions, system events, and plugin execution. Agent Studio supports three types of triggers to cover a wide range of scenarios:

• Conversational Triggers: Activated by user utterances in natural language.
• Event Triggers: Respond to external events via APIs or webhooks.
• Scheduled Triggers: Run plugins at specified intervals or times.

In the following sections, we'll explore each type, including how they work, configuration steps, and examples. For a visual overview, refer to the diagram below illustrating how triggers activate plugins.

graph TD
    A[User Input<br>e.g., Schedule a meeting] -->|Conversational Trigger| B[Plugin Trigger]
    C[External Event<br>e.g., System Alert via API] -->|Event Trigger| B
    D[Scheduled Time<br>e.g., Daily at 9 AM] -->|Scheduled Trigger| B
    B --> E[Moveworks AI Assistant Executes Plugin]
    style A fill:#f9f,stroke:#333,stroke-width:2px
    style C fill:#bbf,stroke:#333,stroke-width:2px
    style D fill:#bfb,stroke:#333,stroke-width:2px
    style B fill:#6B2FC0,stroke:#333,stroke-width:2px,color:#fff
    style E fill:#6B2FC0,stroke:#333,stroke-width:2px,color:#fff

Types of Triggers

Conversational Triggers

Conversational triggers (also known as natural language triggers) enable the Moveworks AI Assistant to detect user intent from phrases or sentences and activate the appropriate plugin. These triggers rely on advanced natural language understanding to match user queries to plugin capabilities.

How they work

When a user interacts with the AI Assistant, the system analyzes the utterance using the Agentic Reasoning Engine. It filters potential plugins based on pre-configured positive and negative examples, rewrites descriptions for optimal matching, and invokes the plugin if there's a strong fit. Configuration involves categorizing generated utterances as positive (should trigger), negative (should not trigger), or skipped.

Tip: Think about natural language—include questions, statements, or commands to make triggering more intuitive

Examples

For a Calendar Management plugin:

• Positive utterance: "Schedule a meeting with the team next week." (Triggers the plugin to initiate booking.)
• Negative utterance: "What's the weather like today?" (Does not trigger, as it's unrelated.)

This ensures the plugin only activates for relevant requests, such as booking or querying events.

How to write good examples

You should:

• Add triggering examples that span a few diverse examples.
• Never add the same / semantically identical triggering examples as both positive or negative examples for the same plugin.
• Do not add positive examples for vague or generic utterances such as "ticket" or "comment".

Additionally, depending on your situation you should either add positive or negative triggering examples.

Event Triggers

Event triggers allow plugins to respond proactively to external events, such as system notifications or API calls, without requiring user input. They are ideal for real-time automation driven by webhooks or integrations.

How They Work

Event triggers use webhooks to receive data from external sources. You create a webhook URL and connect it to plugins. When an event occurs (e.g., a system alert), the trigger sends messages or starts plugins, passing context data for personalized execution.

Examples

For an IT Monitoring plugin:

• Event: A server CPU usage exceeds 80% (sent via webhook).
• Response: The trigger activates the plugin to notify admins and initiate diagnostics automatically.

This enables ambient automation, like alerting teams to outages before users report them.

Scheduled Triggers

Scheduled triggers execute plugins at predefined times or intervals, perfect for routine tasks like reports or maintenance. They ensure consistent automation without manual intervention.

How They Work

These triggers use cron-like expressions or simple schedules (e.g., daily, weekly) to define execution times. When the schedule matches the current time, the trigger invokes the linked plugin, processing any batched data or queries.

Examples

For a Reporting plugin:

• Schedule: Daily at 7:00 AM.
• Action: The trigger activates the plugin to generate and email a sales summary from the previous day.

This is useful for ongoing processes, such as backing up data or sending reminders.