DCT
2:25-cv-01037
Celonis Se v. SAP Se
Key Events
Complaint
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Celonis SE (Germany) and Celonis, Inc. (Delaware)
- Defendant: SAP Se (Germany) and SAP America, Inc. (Delaware)
- Plaintiff’s Counsel: Hogan Lovells US LLP; The Dacus Firm, P.C.
- Case Identification: 2:25-cv-01037, E.D. Tex., 10/10/2025
- Venue Allegations: Venue is alleged to be proper in the Eastern District of Texas because Defendant SAP maintains a regular and established place of business in Plano, Texas, and has allegedly committed acts of infringement within the district.
- Core Dispute: Plaintiff alleges that Defendant’s SAP Signavio Process Transformation Suite infringes four patents related to methods for analyzing business process data, creating unified process protocols from multiple data sources, and generating event logs.
- Technical Context: The dispute is in the field of process mining software, which analyzes data from enterprise systems (e.g., ERP, CRM) to visualize, discover, and improve real-world business processes.
- Key Procedural History: The complaint alleges that the parties had a nine-year partnership, marketed as "SAP Process Mining by Celonis," which ended in 2021 after SAP acquired Signavio, a competitor to Celonis. The complaint also alleges that SAP had pre-suit knowledge of one of the asserted patents (the ’895 Patent) as of March 2024, based on events during the prosecution of one of SAP’s own patent applications.
Case Timeline
| Date | Event |
|---|---|
| 2011-01-01 | Celonis founded |
| 2012-01-01 | Celonis joins SAP’s Startup Focus program |
| 2013-01-01 | Celonis launches its first process mining tool, Celonis 3.0 |
| 2015-01-01 | SAP launches S/4HANA |
| 2015-09-04 | ’861 Patent Priority Date |
| 2016-01-01 | Celonis launches Celonis 4.0 |
| 2016-01-01 | Signavio begins building its process mining software |
| 2016-01-26 | ’257 Patent Priority Date |
| 2016-01-26 | ’895 Patent Priority Date |
| 2018-09-11 | U.S. Patent No. 10,073,895 issues |
| 2018-12-25 | U.S. Patent No. 10,162,861 issues |
| 2019-01-24 | ’708 Patent Priority Date |
| 2020-10-06 | U.S. Patent No. 10,796,257 issues |
| 2021-01-01 | SAP acquires Signavio, ending its partnership with Celonis |
| 2022-02-22 | U.S. Patent No. 11,256,708 issues |
| 2024-03-25 | Alleged date of SAP’s pre-suit knowledge of the ’895 Patent |
| 2025-10-10 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 10,162,861 - “Method for the Analysis of Processes”
Issued December 25, 2018
The Invention Explained
- Problem Addressed: The patent describes prior art methods for analyzing business process data from IT systems as having restricted "flexibility, the performance, as well as the possibility of analyzing arbitrarily complex, a potentially parallel processes" ('861 Patent, col. 1:30-33; Compl. ¶27).
- The Patented Solution: The invention provides a method for storing process data in a specific, sorted data structure ("process protocol") within a computer’s main memory, which allows for efficient analysis ('861 Patent, col. 9:30-46). This method uses a specialized query language, termed an "Advanced Process Algebra Execution" (APE) statement, to receive queries, reconstruct processes from the stored data, and apply "process operators" to perform analyses like filtering or calculating throughput times ('861 Patent, col. 3:1-24, 4:1-12).
- Technical Importance: The claimed approach is described as enabling a "simpler, improved, more flexible, and efficient analysis" of very large and complex data sets compared to prior methods ('861 Patent, col. 1:40-43).
Key Claims at a Glance
- The complaint asserts at least independent Claim 1 (Compl. ¶46).
- Claim 1 recites a computer-implemented method for storage and analysis of process data, comprising the essential steps of:
- storing process data according to a predetermined data structure with at least three attributes: a unique process instance identification, a process step identification, and the sequence of process steps;
- sorting the process steps and storing them at adjacent addresses in memory;
- receiving an "APE statement" defining a query with at least one "process operator";
- executing the APE statement, which involves reading and reconstructing processes from the storage means;
- applying the process operator to the reconstructed processes, which executes a predetermined program code; and
- providing the result for further processing.
U.S. Patent No. 11,256,708 - “Method of Creating Process Protocols”
Issued February 22, 2022
The Invention Explained
- Problem Addressed: The patent addresses the technical challenge of analyzing an "overall process" when its constituent "partial processes are carried out on different computer systems," which consequently creates data on different systems ('708 Patent, col. 1:43-46; Compl. ¶31).
- The Patented Solution: The invention proposes a method where a local computer system imports data tables from at least two different external computer systems. It then "reconstructs the relationships" between these disparate tables, selects primary ("first") and secondary ("second") data tables representing objects and process steps, and creates an "executable transformation." This transformation is then executed to extract data from the original external systems to generate a unified "process protocol" for analysis ('708 Patent, Abstract, col. 2:11-67).
- Technical Importance: This method provides a technical solution for creating a unified analytical data model from physically and logically separate data sources, enabling holistic process analysis where it was previously not possible ('708 Patent, col. 2:5-8).
Key Claims at a Glance
- The complaint asserts at least independent Claim 1 (Compl. ¶95).
- Claim 1 recites a computer-implemented method of creating a process protocol in a local computer system from raw data stored in at least a first and a second external computer system, comprising the steps of:
- (a) importing data tables from the external systems into the local system;
- (b) in the local system, determining and reconstructing the relationships between the imported data tables;
- (c) selecting a "first data table" representing an object for which the protocol is to be created;
- (d) selecting "second data tables" representing process steps;
- (e) determining activities for the objects based on the selected tables and their relationships;
- (f) creating an "executable transformation" adapted to extract data based on the determined activities; and
- (g) executing the transformation to extract data from the external systems, storing it as a process protocol, and providing it to a Process Mining System.
U.S. Patent No. 10,796,257 - “Method for Providing Business Process Analyses”
Issued October 6, 2020
- Technology Synopsis: The patent addresses the problem that non-technical users cannot easily perform business process analyses because IT systems are not prepared to immediately provide a usable event log ('257 Patent, col. 1:39-53). The solution is a method to "provide analytics packages to answer business related process questions" by using a "process sensor" to generate an event log, providing auxiliary data and a data model, and creating executable program modules for graphical analysis based on that model ('257 Patent, col. 1:57-60).
- Asserted Claims: At least independent Claim 1 (Compl. ¶146).
- Accused Features: The complaint alleges that SAP's "Value Accelerators," which provide "ready to use dashboards and over 200 pre-built metrics," and its "investigations" and "dashboards" for data visualization function as the claimed "analytics package" (Compl. ¶¶ 184-188). The complaint includes a screenshot of the SAP Signavio Value Accelerator Library (Compl. p. 67).
U.S. Patent No. 10,073,895 - “Method for Generating an Event Log”
Issued September 11, 2018
- Technology Synopsis: The patent identifies the problem that an "event log can[not] be retrieved immediately from the raw data comprising the traces of executed processes" in conventional IT systems ('895 Patent, col. 1:42-45). The invention provides a solution using a "process sensor" that is executed by a processor to derive process data from raw data, generate unique identifiers and an order for process steps, and store the generated data as an event log in a predetermined structure ('895 Patent, Abstract).
- Asserted Claims: At least independent Claim 1 (Compl. ¶220).
- Accused Features: The complaint alleges that SAP Signavio's data pipelines, which use SQL transformation scripts to "map extracted source system data to business objects" and generate event logs, function as the claimed "process sensor" (Compl. ¶¶ 231, 233, 236). A diagram from SAP's documentation shows data extraction and transformation to illustrate this process (Compl. p. 88).
III. The Accused Instrumentality
Product Identification
The complaint identifies the accused products as the "SAP Signavio Process Transformation Suite" and all reasonably similar products (Compl. ¶23). This suite is alleged to include components such as SAP Signavio Process Insights, SAP Signavio Process Intelligence, and the SAP Signavio Analytics Language ("SIGNAL") (Compl. ¶23).
Functionality and Market Context
The accused suite is described as software for process mining and analysis that integrates with customers' existing IT systems, including SAP ERP systems (Compl. ¶¶ 23, 52). Its alleged functionality includes performing "data extractions and transformations from external source systems" and loading the data into SAP Signavio Process Intelligence for analysis (Compl. ¶53). The complaint alleges that after a nine-year partnership with Celonis, SAP acquired Signavio in 2021 and integrated it with its ERP system to offer an in-house process mining solution in competition with Celonis (Compl. ¶¶ 21-23).
IV. Analysis of Infringement Allegations
’861 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| storing the process data according to a predetermined data structure as process protocol in a storage means... | The accused products store event logs separately in the Apache Arrow columnar format, which allows for fast read access. | ¶57 | col. 9:20-24 |
| wherein the predetermined data structure comprises at least: a first attribute...unique identification... a second attribute...identification of the respective process step... a third attribute...sequence of the process steps... | The SIGNAL language represents an event log where each "case" has a required case id attribute (first attribute), each "event" has a required event name (second attribute), and events are ordered by their end time timestamps (third attribute). |
¶¶63-64 | col. 9:25-33 |
| wherein the process steps...are sorted according to the third attribute...and stored in the storage means at adjacent addresses; | The Apache Arrow columnar format used by the accused products allegedly provides "Data adjacency for sequential access (scans)," satisfying the adjacent storage requirement. | ¶¶67-70 | col. 10:2-4 |
| receiving, by the computer system, an APE statement (Advanced Process Algebra Execution), wherein the APE statement defines a query of process instances...and comprises at least one process operator; | The accused products receive queries in SIGNAL, which is described as a "specialized query language for process analysis" that uses operators to retrieve data and perform calculations. | ¶¶72-74 | col. 9:5-10 |
| executing, by the processor, the APE statement...wherein during reading...processes are reconstructed from the process, steps belonging to the process instances; | The accused products perform process mining, which "uses system event logs...to reconstruct how processes actually run in practice." A visual in the complaint shows a SIGNAL query obtaining average cycle time from an event log (Compl. p. 23). | ¶¶76-78 | col. 4:1-12 |
| providing the result of the execute statement for further processing. | The accused products' SIGNAL query language provides results from its operations on the nested data structure for further analysis and visualization. | ¶¶85-87 | col. 4:13-14 |
- Identified Points of Contention:
- Scope Questions: A central question may be whether SAP's "SIGNAL" query language qualifies as an "APE statement" as recited in the claim. The analysis will likely focus on whether SIGNAL's operators and functions are structurally and functionally equivalent to the "process operators" described in the ’861 Patent specification.
- Technical Questions: The claim requires storing sorted process steps at "adjacent addresses." The complaint maps this to the "Data adjacency for sequential access" feature of the Apache Arrow format. A point of contention could be whether this feature of a columnar database format meets the specific "adjacent addresses" limitation as understood in the context of the patent.
’708 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| (a) importing the data tables from the first external computer system and from the second external computer system into the local computer system... | The accused products establish connections to a variety of source systems (e.g., SAP ERP, AWS S3, ServiceNow) and leverage connectors for data transfer. | ¶¶105-106 | col. 2:26-30 |
| (b) in the local computer system, determining and reconstructing the relationships between the imported data tables, | The accused products perform a "Data conversion" step to transform extracted data into cases and event logs, and the data model allows for creating context relationships between data from different tables. A diagram illustrates this data modeling layer (Compl. p. 35). | ¶¶110-114 | col. 2:31-33 |
| (c)...determining and selecting a first data table from the...imported data tables, wherein each data record...represents an object for which the process protocol is to be created, | The accused products can create event logs for specific "business objects," such as purchase orders, which serve as the object for the process protocol. | ¶¶117-118 | col. 2:34-38 |
| (f)...creating an executable transformation...adapted to extract data from the first data table and from the second data tables based on the determined activities, and to store the extracted data as process protocol... | The accused products use a "process data pipeline" to transform and load data. The result of this transformation is an event log that is loaded into the process. | ¶¶129-130 | col. 2:48-58 |
| (g)...executing the transformation, wherein during execution...the data is extracted from the data tables stored in the external computer systems...and providing the created process protocol to a Process Mining System... | The accused products' ETL components are brought together in "Data models" to transform data into an "event-log format" and connect it to an "investigation to start the process analysis." | ¶¶132, 136 | col. 2:59-67 |
- Identified Points of Contention:
- Scope Questions: The claim requires "reconstructing the relationships" between tables imported from different systems. The dispute may turn on the definition of "reconstructing," and whether the automated or user-guided data mapping and transformation within the SAP Signavio suite meets this limitation.
- Technical Questions: A key technical question could be how the accused product determines relationships between tables from distinct external systems that may lack predefined foreign key relationships. The complaint suggests this is part of the "Data conversion" and "Data model" functionality, but the evidentiary basis for this capability may be a focus of the dispute.
V. Key Claim Terms for Construction
For the ’861 Patent
- The Term: "APE statement (Advanced Process Algebra Execution)"
- Context and Importance: This term is central to Claim 1. The infringement case for the ’861 Patent hinges on whether SAP's proprietary "SIGNAL" query language falls within the scope of an "APE statement." Practitioners may focus on this term because the patent uses a specific, non-industry-standard term that it defines internally.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification describes an APE statement more generally as a "query language" that combines "process operators" (e.g., matching, rework, loop detection) with "common data base functions" to query a process protocol directly ('861 Patent, col. 3:5-24). This could support an argument that any query language performing these functions on a similar data structure is an "APE statement."
- Evidence for a Narrower Interpretation: The patent provides a formal definition of an APE statement and its three "base operators": Case Selection, Crop Selection, and Event Projection ('861 Patent, col. 4:40-67). This could support a narrower construction requiring a query language to be built upon these specific foundational operators.
For the ’708 Patent
- The Term: "reconstructing the relationships between the imported data tables"
- Context and Importance: This term defines the core technical step for unifying data from disparate sources. The infringement analysis will depend on whether the data mapping, transformation, and modeling functions in the SAP Signavio suite perform this claimed "reconstruction."
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification suggests this can be done by determining "via which attributes the data tables are related to each other," such as when attributes are "identical or similar" ('708 Patent, col. 3:1-8). It also allows for the use of imported "data models" ('708 Patent, col. 3:16-24). This suggests a flexible, multi-faceted approach rather than a single, rigid method.
- Evidence for a Narrower Interpretation: A defendant could argue that "reconstructing" implies an automated discovery and re-creation of formal database relationships (e.g., foreign keys) that may not exist between tables from separate systems, and that a user-driven mapping process in the accused product does not meet this requirement. The description of using imported data models could be interpreted as a necessary, rather than optional, part of the reconstruction.
VI. Other Allegations
- Indirect Infringement: The complaint alleges inducement of infringement for all asserted patents. The allegations are based on SAP's advertising and encouraging of customers to use the Accused Products for their intended purpose, which allegedly includes performing the patented methods, such as performing "data extractions and transformations from external source systems" (Compl. ¶¶ 89, 140, 213, 266).
- Willful Infringement: The complaint alleges willful infringement for all asserted patents. For all patents, it alleges knowledge at least from the filing of the complaint (Compl. ¶¶ 91, 142, 215, 268). Crucially, for the ’895 Patent, it alleges pre-suit knowledge since at least March 25, 2024, when a Celonis published application (which issued as the ’895 Patent) was allegedly "discussed during prosecution of SAP’s patent application No. 17/836,848" (Compl. ¶¶ 43, 268).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: can the term "APE statement," which is specifically defined within the ’861 Patent, be construed to cover SAP's "SIGNAL" query language? The outcome may depend on whether the term is interpreted functionally based on its purpose or structurally based on its specific disclosed "base operators."
- A second central question will be one of technical implementation: does the SAP Signavio suite’s process of connecting to multiple data sources, allowing user-defined transformations, and generating event logs perform the specific, ordered steps of "reconstructing relationships" and creating an "executable transformation" as required by the ’708 patent, or is there a fundamental mismatch in the operational workflow?
- A key evidentiary question for damages will be one of pre-suit knowledge: does the prosecution history of SAP's own patent application provide sufficient evidence to establish that SAP knew of the ’895 Patent before this lawsuit was filed, thereby opening the door to a finding of willful infringement and potential enhanced damages?