Celonis Context Model
The second layer of our platform sandwich is the Celonis Context Model.
Introduced in May, the Context Model uses process data from across all of your systems, applications, and devices to show you how your operations run. It’s then enriched with your specific business knowledge. “Think, company goals, how you interact with customers and partners, constraints and guardrails,” said Ana. “Which is how you keep your AI on mission and in bounds.”
And with the acquisition of Ikiai Labs, the Context Model has an extra layer of intelligence, enabling you to predict outcomes, simulate scenarios, and get recommendations on what action you should take next.
Over the past few months, we’ve announced multiple updates across this layer of the Celonis Platform, starting with Task Mining.
“For your AI agents to drive workflows you can trust, they need to see how work gets done outside your systems, on people's desktops,” said Ana.
In May, we announced AI-driven Task Discovery in private preview. As of today, it’s generally available.
With AI-driven Task Discovery, Celonis uses AI to cluster millions of desktop clicks into meaningful business tasks, automatically connecting desktop activity to your business objects and end-to-end processes. Data privacy and GDPR are built in, and it can scale to 10,000 users and millions of data points a day.
“And it pays off extremely fast,” said Julian. “For example, a global hospitality company closed critical visibility gaps across their manual Contract-to-Cash process in just two weeks and built a $3.5 million value case,” added Ana.
In addition to AI-driven Task Discovery, a new recommendation tool for Task Mining (built inside Celonis Studio) enables you to understand what steps with your workflows are slow, manual, and repetitive, and what you should do about it. Once your tasks are discovered and clustered, the tool provides a ranked list of the highest-impact opportunities, with recommendations such as:
- "Automate this step,"
- "Put an AI agent here," or
- "Orchestrate this workflow.”
Next, our cohosts moved to the intelligence functions of the Celonis Platform, starting with Annotation Builder.
“You can now write AI-generated annotations back into your data models,” said Ana. “Think, risk scores, categorizations, or extracted entities. No more manual data engineering and no row limit, so it's built for your enterprise.”
In addition, Annotation Builder can now use RAG (Retrieval-Augmented Generation). Instead of judging every case from scratch, the AI looks back at how similar cases were handled before, making predictions more accurate, consistent, and faster.
RAG for Annotation Builder is available in private preview.
Ana and Julian rounded out the highlighted Context Model updates with Celonis Assistant.
“It’s like having a resident Celonis expert on call, helping you unlock value with our platform even faster and easier,” said Julian. “What makes it so powerful is that it’s context-aware about what you are working on screen and has deep platform knowledge.”
You can ask Celonis Assistant questions, and it provides an answer with references to relevant documentation, products, and pre-built solutions. It will even help you prioritize what to fix next by suggesting KPIs and use cases for you to look at.
“We had over 20 customer teams test this out in private preview, and some customers said it cut their troubleshooting and implementation time in half for some tasks,” said Julian. “They called it a total gamechanger for adoption.”
https://delivery-p141552-e1488202.adobeaemcloud.com/adobe/assets/urn:aaid:aem:0470e205-eb7d-4b58-aede-a0f712777e10/original/as/celonis_now_aug_2026_task_mining_ai_task_discovery.png
