Skip to main content

Celonis Product Documentation

Manual Setup Guide (legacy version 2.0 and versions prior to version 2.0)

After finishing this walkthrough (30-45 mins), you will...

  • Have configured and installed Task Mining on your computer.

  • See the analysis with your application usage, desktop processes, and screenshots of each step.

  • Understand how to run a Task Mining project.

  • Be able to adapt the project to your needs.


An existing Celonis Platform team.

Alternately, you can perform a quick set up to only send data to Celonis Platform; see up API key and Data Pool or Task Mining - Configuration Editor for information about the technical details of the data connection.


Before running the Task Mining Setup, please get in touch with the Celonis Servicedesk <> to prepare your Celonis Platform team for Task Mining and enable the required components.

1. Prepare Celonis Platform team for Task Mining
  1. In ML Workbench for Task Mining, go to the examples_taskmining folder. In there, you will find our Jupyter Notebooks.

  2. Edit and follow steps in ML Workbench Notebook A0_TM_initial_set up.ipynb.

  3. After running the notebook, an example config file is generated called config.reconf which can be used in step 2.

2. Configuration
  1. Download the sample configuration file from the ML Workbench `config.reconf.

  2. Download and install the "CelonisTaskMining.Configuration.Installer-2.x.x.msi" file (it's not the Task Mining Client!).

  3. You can adapt the data that is captured (filtering applications, ...) and the data pool connection using the Task Mining Configuration Software which you just installed on your desktop; for details, see Task Mining Desktop Application.

  4. Adapt the Configuration and insert the screenshot bucket ID, first creating a file storage manager bucket (if you don't want to use an existing one) at by clicking Create and entering a name.

  5. Click the bucket and copy the "Bucket-Id" to your clipboard.

  6. Insert the 'Bucket-Id into your config in the Configuration Editor.

  7. Ensure the Application Key you are using has "Create" rights for the Storage Bucket; you can edit the permissions after clicking on the three dots in the Storage Buckets Overview and choosing "Permissions".

3. Start Task Mining


Even if you don't want to run Task Mining for yourself, you can use this step to run a test to check if the configuration works and the data is sent to the Datapool successfully.

  1. Download and install the Task Mining Software ("CelonisTaskMining.Setup.msi").

  2. Start Celonis Task Mining and load the config file; select config.reconfig from the start screen pop-up.

  3. Click Start.


Chrome extension

Required only if:

  • Filtering events by URL in Chrome.

  • Advanced Chrome metadata is needed.


See the Chrome Extension Install Instructions for additional steps.

Extended metadata support for SAP - SAPlogon for Windows

See How to prepare SAP for integration with Task Mining for additional steps.

4. Analyze results


It can take several minutes to send the data to Celonis Platform. If you're curious, you can use SQL to query the most recent data in the Data Integration Transformation using SELECT * FROM "user_interaction_event_log" order by TimestampUTC DESC

  1. Upload the data in the Task Mining software using Options → Shutdown; It will attempt to upload all data before closing the application.

  2. Execute the Notebook `A1_ TM_run_default_pipeline .ipynb` in the ML Workbench.

  3. Investigate results in the example Analysis ("Installation Report Complete") which was installed in step 1.

  4. If you are not seeing any data you can investigate issues by opening the log file of the Task Mining Desktop Software (Options → Open Log Folder) for troubleshooting.


If you want to run this automatically, set up a Schedule to update the data analytics every time there is a data model reload.

Set up the Screenshot Viewer

  • If you enabled screenshots, you can use an Celonis Platform app to inspect them.

  • To install the screenshot viewer, go to the Marketplace -> Search for "Task Mining Screenshots" -> Click Install.

  • Follow the Readme of the Task Mining Screenshot Package to finalize the setup.

5. Adapt the analysis and data jobs to your needs

You may want to, for example:

  • Define custom KPIs, for example, time spent in some specific SAP view.

  • Link Task Mining to existing process mining data.

  • Define a custom start and end of your cases in the notebook B3_TM_generate_data_model.ipynb.

6. IT Deployment on selected agent computers


The same configuration file can be used for several agents. Separate config filles are only required for separate data pools such as for different department or if different application filtering is required.

  1. Define a list of agents.

  2. Provide your IT department with installer software and configuration file as well as the list of agents to install the software.

  3. The installer and configuration file can usually be rolled out centrally by your company's internal IT. The configuration file can be placed in default locations where it is automatically found at the first start of the Task Mining software.

  4. An analysis can be used to monitor the incoming data after a successful set up to ensure everything is working as expected.