Skip to main content

Action Engine Connector

The Action Engine Connector allows you to extract data from the Action Engine to perform usage analyses. You can extract tables such as Skills, Skill Execution, and Signals (including IDs, status, and assignees).

Extracted tables

The following tables can be extracted:

  • Skills (Skill ID, Skill Name)

  • Skill Execution (History of all Skill Executions)

  • Statistic (Basic statistics e.g. #Signals created per Skill)

  • Signals (Signal ID, Skill ID, Signal Description, Status, Assignee User ID, Creation Date, Change Date)

  • Signal Attributes (Signal ID, Sorting, Name, Value)

  • Signal Events (Events per Signal e.g. Event Time, From Status, To Status)

Before you begin

Before you begin the setup, ensure the following requirements are met:

  • Permissions: You must be a Team Administrator to activate the export endpoints in the Action Engine configuration tab.

  • Regional Availability: Endpoints must be available for your realm (currently supported: eu-1 through eu-4, us-1, and us-2).

  • Connector Activation: The "Celonis Action Engine" must be visible as a Cloud Extractor in Data Integration. If it is not visible, contact Support to have it activated for your team URL.

  • Network: If IP-based restrictions are in place, allowlist the standard Celonis IPs.

Configuring the Action Engine connector

You can either create a standalone Data Model or integrate Action Engine data into an existing one.

Option 1: Creating a standalone data model

To create a standalone data model:

  1. Navigate to Data Integration and create a new Data Pool.

  2. Create a New Data Connection:

    • Select Celonis Action Engine.

    • Set the connection type to Direct.

    • Enter an API Token (generated from your user profile).

  3. Create a New Data Job and select the connection you just created.

  4. Add a New Extraction and select all desired tables (Skills, Signals, etc.).

  5. Execute the Data Job: Use Delta Load to preserve historical data. A "Full Load" will delete existing data and replace it with only what is currently in the Action Engine.

  6. Create a New Data Model:

    • Add all extracted tables.

    • Set the Signal table as the Case table and Event as the Activity table.

    • Connect remaining tables via Signal ID and Skill ID.

Option 2: Integrating into an existing data model

To integrate into an existing data model:

  1. Open your existing Data Pool and follow the steps above to create the Data Connection and Extraction. See: Option 1: Creating a standalone data model.

  2. Create a Data Transformation to join the Action Engine data with your process data.

    Example: For a P2P process, create a table (e.g., AE_P2P_SIGNALS) that joins the SIGNAL table with your MANDT, EBELN, and EBELP attributes.

  3. Add the transformation table to your existing Data Model.

  4. Map the connection via the Case Key (e.g., Purchasing Document/Item).

Once the Data Model is loaded, you can begin building analyses to monitor Skill execution history and Signal status changes.

Example Procure-to-Pay transformation script

If you're working on a procure-to-pay process:

DROP TABLE "AE_P2P_SIGNALS";

CREATE TABLE AE_P2P_SIGNALS AS 
SELECT DISTINCT 
    SIGNAL.SIGNAL_ID 
    ,SIGNAL.SKILL_ID
    ,SKILL.NAME
    ,SIGNAL.ASSIGNEE_USER_ID 
    ,SIGNAL.DESCRIPTION  
    ,ATTRIBUTE1.VALUE AS MANDT
    ,ATTRIBUTE2.VALUE AS EBELN
    ,ATTRIBUTE3.VALUE AS EBELP
        --Optional: Add further columns
FROM SIGNAL
JOIN SKILL ON 
    SIGNAL.SKILL_ID = SKILL.SKILL_ID 
LEFT JOIN SIGNAL_ATTRIBUTE AS ATTRIBUTE1 ON 
    ATTRIBUTE1.SIGNAL_ID = SIGNAL.SIGNAL_ID 
    AND ATTRIBUTE1.NAME = 'Client' --Use the exact Attribute name here as used in the Signal
LEFT JOIN SIGNAL_ATTRIBUTE AS ATTRIBUTE2 ON 
    ATTRIBUTE2.SIGNAL_ID = SIGNAL.SIGNAL_ID 
    AND ATTRIBUTE2.NAME = 'Purchasing Document' --Use the exact Attribute name here as used in the Signal
LEFT JOIN SIGNAL_ATTRIBUTE AS ATTRIBUTE3 ON 
    ATTRIBUTE3.SIGNAL_ID = SIGNAL.SIGNAL_ID 
    AND ATTRIBUTE3.NAME = 'Item' --Use the exact Attribute name here as used in the Signal
--Optional: Add further joins
WHERE 
    SKILL.SKILL_ID IN ('') --Add Skill IDs here
GROUP BY 
    SIGNAL.SIGNAL_ID, SIGNAL.SKILL_ID, SIGNAL.ASSIGNEE_USER_ID, SIGNAL.DESCRIPTION ,ATTRIBUTE1.VALUE 
,ATTRIBUTE2.VALUE ,ATTRIBUTE3.VALUE;

SELECT * FROM "AE_P2P_SIGNALS";

Related topics