AI

 I've recently completed an independent research study exploring how Celonis Object-Centric Process Mining (OCPM) can help us understand the execution behavior of AI agents.The idea was simple: If Celonis can provide deep visibility into complex business processes, could we apply the same principles to understand how AI agents execute tasks, interact with tools, and encounter failures?To investigate this, I analyzed execution traces from the WildClawBench dataset using Celonis.Some interesting findings emerged: 720 AI-agent execution runs analyzed, covering 35,771 events. 17,206 tool invocations reconstructed and examined. 1,487 batched tool requests identified, revealing execution structures that are not immediately apparent in conventional event sequences. 1,142 background-process sessions linked to their execution context. Failure-indicator visibility increased from 50.7% to 100% when richer execution metadata was preserved, based on our rule-defined classification framework

Hello Celonis Community,I recently completed an independent empirical study exploring a question that I believe could be relevant to the future of Process Intelligence:Can the principles of process mining be extended beyond conventional enterprise workflows to analyze the execution behavior of autonomous AI agents?Using Celonis and execution traces from WildClawBench V4, I investigated how agent activity sequences, tool interactions, and explicitly flagged execution errors could be reconstructed and analyzed through a process-mining approach.Research scope 720 AI-agent executions 35,771 recorded events 12 AI models 6 task categories Process reconstruction, model-level comparisons, and exploratory early-warning evaluation What did the research find?1. Process observability: AI-agent execution traces could be represented and investigated through a process-mining framework, making execution paths, tool interactions, and flagged errors more interpretable.2. Behavioral differences:

I want to learn Ai in celonis! can i only focus on a specific Ai path ? as Ai is something I'll really like to learn.. Ai is future and i want to upskill myself in Ai plus celonis.. is there need of anything beforehand?

Artificial intelligence is changing the way businesses communicate with customers. From personalizing SMS campaigns to optimizing delivery times and analyzing customer engagement, AI is making SMS marketing smarter and more effective. Platforms like Qalaama help businesses combine reliable bulk SMS services with data-driven strategies to improve customer communication, increase conversions, and automate messaging workflows. What AI features do you think will have the biggest impact on SMS marketing?

We are living in a time where technology is not just evolving — it is accelerating.Artificial Intelligence (AI) is no longer a futuristic concept from movies. It is already transforming the way we live, learn, work, and solve problems.AI is present everywhere:Healthcare uses AI to detect diseases earlyFinance uses AI to predict market risksCompanies use AI to understand customer behaviorSelf-driving cars use AI to make decisions in real timeRobots use AI to interact with the world intelligentlyAI is not just about coding.AI is about thinking.AI is about solving real-world problems.AI is about building systems that can learn, adapt, and improve.As a student passionate about Artificial Intelligence, Machine Learning, and Data Science, I am continuously learning:• Python programming for problem solving• Data Structures and Algorithms for logical thinking• Mathematics for understanding models• Machine Learning for prediction and decision making• Data Science for extracting insights from da

Webinar Sessions | Celonis x AI: 𝗙𝗿𝗼𝗺 𝗕𝘂𝘇𝘇𝘄𝗼𝗿𝗱𝘀 𝘁𝗼 𝗔𝗰𝘁𝗶𝗼𝗻𝗮𝗯𝗹𝗲 𝗦𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝘀AI is trending everywhere but, how do you turn hype into real business impact?Join us to explore:- Showcase of AI capabilities in Celonis, featuring a hands-on demo for service tickets.- Processand capabilities scalable for AI deployment and operation- How to get started and initiate pilot use cases📅 19 September & 26 September | 🕒 10:20 - 11:00 AMSpeakers: Federica Tella (Consultant) & Edoardo Puddu (Data Scientist)𝗦𝗲𝗰𝘂𝗿𝗲 𝘆𝗼𝘂𝗿 𝘀𝗽𝗼𝘁:- 𝟭𝟵𝘁𝗵 𝗦𝗲𝗽𝘁𝗲𝗺𝗯𝗲𝗿 𝟮𝟬𝟮𝟱:https://lnkd.in/duYZ-sQB- 𝟮𝟲𝘁𝗵 𝗦𝗲𝗽𝘁𝗲𝗺𝗯𝗲𝗿 𝟮𝟬𝟮𝟱:https://lnkd.in/dubVmgsf  

Sharing this webinar to the broader community.  It is a 30-minute webinar delivered by Celonis and AkzoNobel.  Link below to register.  GBS Beyond Cost: Working Capital & AI Transformation Webinar | CelonisLearn how Celonis Process Intelligence helps GBS organizations drive AI transformation and improve working capital performance. Register for our webinar on April 28.We’ll explore how to optimize working capital, accelerate AI transformation, and reduce Days Sales Outstanding (DSO). We are thrilled to feature AkzoNobel, who will share their firsthand success story and transformative use case with Celonis.

Can someone help me understand what that error in red means?I’m following the AI annotation builder training and I cant go ahead from here

Dear Celonis Community,in the documentation for the Process CoPilot, there is the section “Documents” for which I can’t find any documentation. https://docs.celonis.com/en/knowledge-input.html Is there any documentation about how to apply this new input of knowledge (e.g. PDFs or images) and how it synergizes with the data from the records/knowledge model? Best regards and many thanks in advanceJan 

Currently, I am doing the “Introduction to Studio Assets” . Everything worked fine, until now. During the creation of a process copilot, described here: https://academy.celonis.com/learn/course/introduction-to-studio-assets/introduction-to-studio-assets/introduction-to-studio-assets?client=partner&page=11 I ran into a problem.I activated the AI, started creating a process copilot (aas described in the training, see link) and finished step 1 “AI instructions”  (selecting the knowledge model and the task as described). Now I’m in step 2 “Knowledge Input”. The training says that I have to select a metric. So I clicked on “add metric” and tried desparately to select one. Now, according to the video, a list of metrics should appear. But this does not happen. I see no list of metrics and therefore I cannot select one! What can I do to finish the “Introduction to Studio Assets”?

I have a strange issue while creating copilot. I want the copilot to read data from one table only. The table contains creation date, closure date, case id. While simple questions like created cases in June are ok for the copilot to handle, wekely or daily trend charts are not produced. The copilot can’t extract week or month no from date columns to produce the trend charts. AI services enabled are: Claude 3.7 Sonnet (AWS)GPT 4o (Azure OpenAI) From time to time the copilot get’s frozen while thinking, it can take up to few hours for it to spin without results. Anyone faces similar issues? 

  Latest trend: #AI 

How can I get the process copilot to use the record attributes saved on a knowledge model to search for the “output” rather than using the sample from the prompt? I have tried removing the sample from the prompt and it does not work. I have used the attributes on the knowledge input section of the process copilot as well and it also seems not to be enough.

Hi everyone I have a challenge that I’m trying to figure out related to pycelonis.I’m currently trying to get the component values from a view using pycelonis. Unfortunately, I couldn’t find any information inside the documentation that can help me with this.It seems that the documentation only explains how to work with analysis or knowledge models. Does anyone know how to work with components of a view inside pycelonis?Best wishes!

I have generated a zip file with the content-cli module and now I need to download it locally, but every time I press the “Download” button it creates an HTML file. That HTML redirects me to a “Secure download” which re-downloads the same HTML. How can I avoid the HTML and just retrieve the zip file?

Hi everyone,I want to use the Annotation Builder, but I don’t have a unique ID.For example, the same material number exists in two different business units as you see in the picture.What’s the best way to handle this? Thanks for your help!  

I am unable to get the option to schedule a notebook in my ML workbench.Does it require admin access to get the option?  Documentation says to have a valid executable file, which I have. Any code to be written in the notebook to make it eligible for scheduling?  

Hi all,I want to use the Process CoPilot with the built-in python tool in order to execute ad hoc calculations.https://docs.celonis.com/en/tool-activation.html#UUID-3c37cb75-0631-254f-a5b1-1a70ffd8d5f4_UUID-5523d1d0-d4b7-e2d3-4ebe-f999dffd6357and https://docs.celonis.com/en/tool-activation.html#UUID-3c37cb75-0631-254f-a5b1-1a70ffd8d5f4_UUID-b801ca95-08b6-fe38-c5c5-89bffe3f22aaI don’t understand the examples...My question:HOW can I implement this? HOW is the data going to be accessed and loaded in order to generate a dataframe? The examples in the docs don’t help me at all … Why is a df generated from a “KPI”? There is no relation between “get_data” output and the “code” …  HOW should I use get_data and python? HOW do these tools interact?Thank you in advance and best regardsJan  

Dear Celopeers, Currently I am trying to extend the simulation capacity of Celonis in the Machine Learning workbench. To do so, I need to import an event log through, I think, PyCelonis. I would like to use the UI of the Process Explorer in a premade analysis of Studio to define the activities and connections that are imported.The following link suggests this used to be possible within PyCelonis 1.7 using the ‘Share your selection’ URL. (Analysis: Pull data - PyCelonis) This appears like a very elegant and user friendly solution to me.However, I can't figure out how to do so using the newer PyCelonis 2.12. I did find out how to navigate to the correct space, package, and analysis through different IDs. But the Python methods attached to the analysis object created through .get_analysis() do not make apparent how to perform the filtered data extraction I alluded to above.Any ideas or references on how to do this? Thank you in advance, Bas

Hi all, is it possible to code in the Academy Studio or not. For me it always show that there is no data model or data pool that i can use. Thanks for your helpbest regards Louis

I am working on a script in PyCelonis to search assets and compile a list of the data models/Knowledge models that the assets are using, with the aim of non manually figuring out which DM/KMs are currently in use by production assets.  Using PyCelonis I have been able to navigate down to the asset level and find lots of information for each asset, but have so far not found anywhere that you can identify the DM/KM an asset is built on. The “.dict()” function provides a lot of information and I thought would be the intuitive place to include this kind of information but it is nowhere to be found, except when the asset is a view and the KM gets referenced in the yaml code that is provided in “.dict()” Does anyone know if this is information that we can pull from PyCelonis? And, if so, how to find this information on the asset level? Thanks in advance for your help!

Tried with Python script to extract table from the web page, however it didn't work.

I just completed the "Comparative and Predictive Process Mining" course to learn more about ML capabilities, prediction, simulation, etc. I made an assumption that the example notebook/code that was shown in the video regarding feature recommendation and model training would be OOTB or readily available. I cannot find any examples on Celonis site or web. Are there examples for use or is it on us(me) to develop from scratch?

Hi Team,Im trying to convert the data i need from the datapool into a pandas dataframe for further calculation.In the documentation i could only find how to read the parquet files and convert into Pandas.Is there any way?Let me knowThanks 

Hi all,I want to extract data from a table in the schema explorer, which is loaded via an extraction, the standard way is to do it is by loading the table into datamodel and pulling the data from the datamodel using pycelonis as mentioned in the pycelonis documentation I am curious to know if there is any other way to do it without loading the table into datamodel (via the schema explorer if possible)