LICENSEWARE

Celonis AI, automation and add-on licensing

This article is about the optional and AI-related parts of a Celonis subscription: AI Features and the AI Addendum, Bring Your Own Model, Action Flows and the Orchestration Engine, Make Enterprise credits, APIs and fair use policies, and Bring Your Own Key.

On This Page

Beyond process modules and capacity, a Celonis subscription often includes AI-powered features, automation engines and integration interfaces. Each has its own metric or policy. This article follows the Product Description (September 2026), the Metrics Descriptions, the Artificial Intelligence Addendum (July 2025), the Additional Services Addendum (September 2024) and the Make Enterprise documents.

AI Features and the AI Addendum

The MSA defines an AI Feature as a product, feature or functionality in the Celonis Services that uses artificial intelligence, machine learning, large language models or similar technologies, and says the Artificial Intelligence Addendum forms part of the agreement to the extent AI Features are enabled in the customer’s environment.[1] The AI Addendum prevails over the agreement in a conflict, but only for AI Features, and applies on use or enabling of the features.[2]

Key terms:

  • Ownership. Celonis keeps all rights in the AI Features; as between the parties the customer keeps rights in its AI Input and AI Output. The addendum acknowledges that AI Output may be similar or identical to output given to other users, and that Celonis uses AI Providers (third-party foundation model providers named in its sub-processor list) so that AI Input may be used or accessed outside Celonis’s hosted environment.[2]
  • Use of data. Celonis may process AI Input and usage information to provide the features, meter consumption and ensure compliance with Order limits, and to improve its products. It states that it will not permit AI Input to be used to create, train or improve third-party LLMs except to improve the output for the customer’s sole benefit.[2]
  • Restrictions. The customer may not use the AI Features to infringe third-party rights, to develop, train or improve any LLMs unless authorised, to represent AI Output as vetted by Celonis or as wholly human-generated, or for automated decision-making with legal or similarly significant effects without adequate human review.[2] The customer acknowledges that the features are not a high-risk AI system as defined in the EU AI Act and is responsible for ensuring they are not used for purposes listed in Article 6 or prohibited by that Act.[2]
  • Liability. Breach of these obligations can lead to termination or suspension, and the liability caps in the agreement do not apply to breach of the customer’s obligations in the section on access and acceptable use.[2] A customer rolling out AI features to many users should note that this exposure is not limited by the agreement’s cap.
  • Disclaimer. Celonis disclaims responsibility for the underlying LLMs and their training data, and the customer is solely responsible for evaluating AI Output for accuracy and bias. Celonis passes through any intellectual property commitment it receives from an AI Provider, subject to the customer complying with the provider’s required mitigations.[2]

The general topic is covered in generative AI licensing.

Metering: AI Outputs

AI consumption is counted in AI Outputs: one Process Copilot response, one record processed and annotated by the Annotation Builder, or one Chat API response. Usage by Insights Explorer and AI Assistants and usage on a customer’s own LLM is excluded.[3] Quotas are annual and unused AI Outputs do not roll over.[3] See Celonis platform capacity for the entitlement ratios.

Bring Your Own Model

Bring Your Own Model (BYOM) integrates customer-managed LLMs as the underlying engine for Celonis generative AI applications such as Process Copilots and the AI Annotation Builder. In the Product Description it is an optional add-on in the Infrastructure edition and included in the AI Platform edition.[4] Under the AI Addendum the customer keeps title to its LLM; Celonis provides reasonable assistance, but the customer is solely responsible for compatibility, integration cost, legal compliance, security and management of its LLM, and must stop using it with the services and remove it from Celonis systems at termination.[2] The AI Innovation offering in the OCPM description bundled AI Assistants, Insights Explorer, Annotation Builder, Process Copilot, Chat API, BYOM and early access to the Tools API, with usage following the AI Output limits on the Order.[5]

Action Flows, Orchestration Engine and Agent Tools

Action Flows are a low-code automation engine integrated with the platform, included in both editions in the Product Description.[4] Their usage is measured in Executions and Operations: an Execution occurs each time a record is created, updated or deleted in an operational system, and each trigger run or action is an Operation.[5]

The Orchestration Engine lets customers digitise processes with a low-code builder and human-in-the-loop interactions. It is an optional add-on in the Infrastructure edition and included in the AI Platform edition.[4] In the OCPM description it is priced by the number of Orchestration Use Cases, Processes or Domains. An Orchestration Engine Use Case is a defined collection of one or more Process Orchestrations grouped for a measurable business objective, with a maximum of 15 Process Orchestrations per use case, and use cases may be deployed only for processes the customer has licensed.[5] If several use cases are combined into a composite orchestration, each is counted individually towards the entitlement: a composite spanning five use cases counts as five.[5] Buying the Orchestration Engine licence grants an entitlement to an unlimited number of Action Flows.[5]

Agent Tools (MCP), formerly the MCP Server, exposes Celonis Process Intelligence to AI agents through the Model Context Protocol or an OpenAPI specification. It is an optional add-on in the Infrastructure edition and included in the AI Platform edition.[4] The Knowledge Model API, formerly the Intelligence API, is a set of read-only REST APIs for consuming knowledge model information in external platforms, and the Chat API, which deploys a Process Copilot into systems such as Microsoft Teams or Slack, is subject to Knowledge Model API access and limits.[4] The Celonis Connector for Power BI is an optional add-on, priced in the OCPM description by the number of Process Modules purchased.[4][5]

Bring Your Own Key

Bring Your Own Key (BYOK) is an optional add-on that lets customers hold exclusive ownership and management of the encryption keys that protect data at rest. By opting in the customer acknowledges that BYOK renders data unreadable even to Celonis and that losing the key means losing access to the data entirely, with no Celonis liability or obligation to recover it. BYOK may not apply to all deployment environments or features, such as Action Flows, in which case standard Celonis-managed keys are used.[4] Premier Plus Support includes BYOK on Public Cloud.[6]

APIs and pipeline tools

The Additional Services Addendum covers optional items such as APIs, the Celonis Schema, Early Adopter Features, Non-Celonis Applications and Proofs of Value. Use of any Celonis API is subject to the API Fair Use Policy.[7] Under that policy, access credentials may be shared only with the customer’s employees or independent contractors; the licence to use the APIs to build integrations is limited, non-transferable and revocable and runs alongside the subscription term; and Celonis may limit any use levels that exceed the usage capacity or frequency designated for the API access purchased.[8] The Data Pipeline Orchestration Tools Fair Use Policy grants a similar limited licence to use and modify the Tools to build integrations, bars competitors of Celonis from using the Tools or using them to develop a competing product, and caps Celonis’s liability at USD 1,000 where exclusion is prohibited.[9]

Non-Celonis Applications, third-party software interoperating with the services and possibly acquired via the Marketplace, are subject to the agreement and any applicable third-party terms; where the customer has a separate agreement with the third party, the Celonis agreement does not apply to that application.[7] Additional Services are provided as-is, with no service level agreement or support, and Celonis’s total liability for them is capped at USD 1,000 where exclusion is prohibited.[7]

Make Enterprise by Celonis

Celonis also sells Make Enterprise, a no-code workflow builder for connecting apps and APIs. It is metered in Credits. A Credit is the billing unit consumed when executing a scenario. Each time a trigger runs it consumes one Credit regardless of how many bundles it retrieves, and each action performed consumes one Credit. AI modules on Make’s own AI provider may be usage-based, such as token consumption, while AI modules using the customer’s own API key consume a flat one Credit per Operation, with token usage billed by the customer’s provider.[10] The total Credits depend on the Order.[10]

The product description sets a usage allowance relative to Credits: every 120,000 licensed Credits a year includes 60 GB a year of data transfer, 10 MB of internal data storage and 10 MB of incomplete executions storage, with a webhook queue of 667 up to 10,000. Up to 1,000 data stores per organisation are allowed regardless of Credits. Feature limits include a one-minute minimum interval, a 40-minute scenario timeout, a 1 GB maximum file size, 1,000 Make API calls a minute and 60 days of execution log storage.[10]

The Make Order Supplement modifies the MSA for Make Services. Use is for internal purposes unless the Order includes an external use licence for OEM offerings, which permits providing Make functionality embedded in the customer’s product to third-party subscribers on a hosted subscription basis only, with no sublicensing without written agreement, and makes the customer liable for those subscribers’ acts.[11] The supplement also bars publicly disseminating performance information such as benchmarks.[11]

Practical checks

  • Confirm which AI Features are enabled and whether usage is on Celonis-provided models or the customer’s own, since BYOM usage is excluded from AI Output counts.[3]
  • Count Orchestration Engine use cases individually, including inside composite orchestrations.
  • Review who holds API credentials; only employees and independent contractors may.
  • For Make, compare Credit consumption in the Make dashboard with the Order.

Out of scope

This article does not cover pricing, partner marketplace apps, or the Make-specific privacy and data processing documents. For the capacity units see Celonis platform capacity.

References

  1. Celonis Master Services Agreement (June 2026)Dated "version June 2026".Retrieved 2026-10-08.
  2. Celonis Artificial Intelligence Addendum (July 2025)Dated July 2025.Retrieved 2026-10-08.
  3. Celonis Metrics Description for Celonis Process Intelligence - AI Platform (May 2026)Dated May 2026.Retrieved 2026-10-08.
  4. Celonis Product Description (September 2026)Footer "Last updated: September 2026".Retrieved 2026-10-08.
  5. Celonis Metrics Description for OCPM Infrastructure (November 2025)Dated November 2025; archived prior version.Retrieved 2026-10-08.
  6. Celonis Premier Plus Support Addendum (May 2026)Dated May 2026.Retrieved 2026-10-08.
  7. Celonis Additional Services Addendum (September 2024)Dated September 2024.Retrieved 2026-10-08.
  8. Celonis API Fair Use PolicyUndated.Retrieved 2026-10-08.
  9. Data Pipeline Orchestration Tools Fair Use Policy (April 2025)Dated April 2025.Retrieved 2026-10-08.
  10. Make Enterprise by Celonis Product Description (August 2025)File name August 2025.Retrieved 2026-10-08.
  11. Order Supplement for Make Services (August 2023)Dated August 2023.Retrieved 2026-10-08.

See also

Esc