Skip to main content

Action Engine - Skills

Skills Deprecation

Effective August 1st 2025, Skills features can no longer be purchased as part of a Celonis Platform license. Celonis continues to maintain Skills (by fixing bugs and ensuring that performance remains strong) but no further feature development will take place.

To continue working with your process improvement and automation use cases, we suggest using our Action Flows, Views, and/or Orchestration Engine features.

A Skill is the automated logic that monitors your processes for inefficiencies. It acts as a bridge between data insights and operational action, transforming identified process gaps into manageable tasks for your business users.

Creating a Skill is more than just writing a query; it is a four-step process of identifying, routing, and automating process improvements. Understanding this lifecycle ensures that the Signals generated provide maximum value to business users.

  1. Detection (Logic): The lifecycle begins with Signal Configuration. You define the specific process anomaly you want to target using PQL (Process Query Language).

    • Example: "Identify all invoices where the 'Payment Terms' have been changed manually."

  2. Routing (Assignment): Once an anomaly is detected, the Action Engine must determine who is responsible for fixing it. This is handled via User Routing. You can route based on:

    • Static Assignment: A specific user or group.

    • Dynamic Assignment: Using data columns (e.g., the "Purchasing Group" column in your data model).

  3. Execution (Automation): A Skill becomes powerful when it offers a resolution. By adding an Action configuration, you allow the user to fix the issue directly from their Inbox.

    • Manual Actions: Opening a link to a specific SAP transaction.

    • Automated Actions: Triggering a Workflow to update a system automatically.

  4. Feedback and Optimization: The final stage is the feedback loop. When a business user acts on a Signal in My Inbox, their feedback is captured. Analysts use this data to refine the detection logic and reduce "noise" (false positives).

Related topics