Adaptive Classification Tech LLC v. Open Text Inc

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Adaptive Classification Technologies LLC (Texas)
  • Defendant: Open Text Inc. (Delaware)
  • Plaintiff's Counsel: Bradford Black P.C.
• Case Identification: 1:26-cv-00866, W.D. Tex., 04/07/2026
• Venue Allegations: Plaintiff alleges venue is proper in the Western District of Texas because Defendant Open Text Inc. maintains a regular and established place of business in the district and has committed acts of infringement there.
• Core Dispute: Plaintiff alleges that Defendant's Axcelerate eDiscovery Platform infringes a patent related to systems and methods for determining when to terminate a technology-assisted document review process.
• Technical Context: The technology at issue is in the field of electronic discovery ("e-discovery"), specifically concerning machine learning protocols known as Technology-Assisted Review ("TAR") used to classify massive volumes of electronic documents.
• Key Procedural History: The complaint alleges that Plaintiff provided Defendant with notice of the patent-in-suit and its alleged infringement via a letter dated March 26, 2026, approximately two weeks before the complaint was filed.

Case Timeline

Date	Event
2015-06-19	Priority Date for U.S. Patent No. 10,445,374
2019-10-15	U.S. Patent No. 10,445,374 Issued
2026-03-26	Plaintiff sends notice letter to Defendant regarding alleged infringement
2026-04-07	Complaint Filed

II. Technology and Patent(s)-in-Suit Analysis

• Patent Identification: U.S. Patent No. 10,445,374, "Systems and Methods for Conducting and Terminating a Technology-Assisted Review," issued October 15, 2019 (the "'374 Patent").

U.S. Patent No. 10,445,374 - "Systems and Methods for Conducting and Terminating a Technology-Assisted Review"

The Invention Explained

• Problem Addressed: In large-scale document reviews, a key challenge is determining when to stop the review process while ensuring that a sufficient number of relevant documents have been found (i.e., achieving "high recall") '374 Patent, col. 3:26-31 Traditional methods for measuring recall can be problematic due to imprecision, bias, and the difficulty of defining an absolute threshold for "high recall" '374 Patent, col. 3:39-64
• The Patented Solution: The invention proposes a method to terminate a TAR process with a statistical guarantee of reliability '374 Patent, col. 8:46-49 The process involves first identifying a "target set" of known relevant documents '374 Patent, col. 6:31-34 Then, a separate, "independent" classification process (like a TAR workflow) is run on the entire document collection '374 Patent, col. 7:20-25 The system terminates the review based on a comparison between the results of this independent process and a characteristic of the pre-defined target set, ensuring that a target level of recall is achieved with a certain probability '374 Patent, abstract '374 Patent, col. 4:28-31 FIG. 1 of the patent illustrates this workflow, showing the initial identification of a target set (1020), the use of an independent search (1040), and the determination of a stopping criteria based on a comparison (1060).
• Technical Importance: The method provides a structured, provably reliable framework for ending a TAR process, moving beyond subjective reviewer decisions or less reliable sampling methods to offer a defensible endpoint for e-discovery reviews '374 Patent, col. 6:56-60

Key Claims at a Glance

• The complaint asserts at least independent claim 1 Compl. ¶54
• Essential Elements of Claim 1:
  • A system with a processor and memory storing instructions.
  • The system receives an identification of a "target set of documents" from a larger collection.
  • It executes a classification process that uses a "second iterative search strategy" to classify documents in the collection, which "does not distinguish" between documents in the target set and other documents in the collection.
  • The process enables training of a classifier "using documents in the target set."
  • It terminates the classification process based on a "comparison" between the results of the second search strategy and a characteristic of the target set.
  • This termination ensures the process achieves a "target level of recall with a certain probability."
• The complaint does not explicitly reserve the right to assert dependent claims.

III. The Accused Instrumentality

Product Identification

• The OpenText Axcelerate eDiscovery Platform, including its Axcelerate TAR 2.0, Predictive Search, and continuous active learning ("CAL") functionalities Compl. ¶47

Functionality and Market Context

• The accused platform is a TAR system designed to help users classify large document collections for e-discovery Compl. ¶48 It employs a "continuous active learning" (CAL) protocol, which the complaint alleges "learns from all reviewer coding decisions in real time to deliver the most relevant results" Compl. ¶48 Compl. ¶51 The process can be initiated with a "seed set" of documents identified by the user as relevant, which helps train the algorithm Compl. ¶50 The complaint includes a screenshot of the "Predictive Search" feature, which it alleges compares known relevant documents against the entire corpus. Compl. p. 12
• The complaint alleges that the platform's workflow allows a user to "continue until the desired recall rate is reached" and to track progress to determine "when it is time to stop" Compl. ¶52 Compl. p. 15

IV. Analysis of Infringement Allegations

'374 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
receive an identification of a target set of documents in a document collection...	The system allows a user to initiate a TAR 2.0 protocol by providing initial "seed documents" found to be relevant, which are fed to the system to help train the algorithm.	¶50	col. 6:63-66
execute the classification process that enables training of a classifier using documents in the target set, wherein the classification process utilizes a second iterative search strategy...	The Axcelerate TAR 2.0 platform executes a CAL process where the algorithm "learns and improves continuously throughout the review process" and "every review decision...is used to train and improve the algorithm."	¶51	col. 7:26-29
...which does not distinguish between documents in the target set and documents in the document collection to classify documents in the document collection...	The CAL algorithm allegedly "continuously ranks the entire document collection," applying its classification criteria across the whole corpus.	¶51	col. 18:35-39
...and terminate the classification process based upon a comparison between the results of the second search strategy and a characteristic of the target set of documents...	The platform enables users to terminate the process when a desired recall objective is met, and its "Predictive Search" feature allegedly compares known relevant documents against the entire corpus to assess if the final set is inclusive of all relevant content.	¶52	col. 4:28-31
...wherein the classification process achieves a target level of recall with a certain probability upon termination.	The complaint alleges the platform's workflow terminates when a "desired recall rate is reached" and that success is measured by balancing recall and precision, with a target recall level (e.g., 80 percent) as a common standard.	¶52	col. 4:15-17

Identified Points of Contention

• Scope Questions: A central question may be whether the user-provided "initial seed documents" in the accused system Compl. ¶50 function as the claimed "target set of documents." The defense may argue that a seed set used for initial training is functionally different from the patent's "target set," which appears to be used as a benchmark for a termination test.
• Technical Questions: The analysis may focus on what constitutes the claimed "comparison." The complaint alleges termination occurs when a "desired recall objective has been reached" Compl. ¶52 The court may need to determine if tracking progress toward a recall goal is technically equivalent to the claimed "comparison between the results of the second search strategy and a characteristic of the target set." Another point of contention could be the meaning of "does not distinguish," as the accused CAL algorithm necessarily uses the coding of the seed/target set to inform its ranking of the entire collection.

V. Key Claim Terms for Construction

Term: "target set of documents"

• Context and Importance: The definition of this term is fundamental to the infringement theory. The complaint equates a user's initial "seed set" with the claimed "target set." The case may turn on whether this interpretation is supported by the patent's intrinsic evidence.
• Intrinsic Evidence for Interpretation:
  • Evidence for a Broader Interpretation: The claim language states the target set consists of "documents identified as relevant as part of a first search strategy" '374 Patent, col. 18:29-32, which could be broadly interpreted to include any pre-identified relevant documents, such as a seed set. The specification also notes that documents can be added to the target set if a "reviewer indicates the document is 'relevant'" '374 Patent, col. 7:1-3, a process similar to creating a seed set.
  • Evidence for a Narrower Interpretation: The patent frequently describes the target set in the context of random sampling from the collection '374 Patent, col. 4:31-34, which may suggest a more formally constructed, statistically significant set rather than an ad-hoc seed set. The patent also contrasts its "target technique" with a "control set," emphasizing its role in providing an "unbiased measurement of recall" '374 Patent, col. 6:56-60, potentially narrowing its scope to a set created specifically for this purpose.

Term: "terminate the classification process based upon a comparison"

• Context and Importance: This term defines the trigger for stopping the review. Infringement depends on whether the accused platform's method for determining completion-allegedly by reaching a "desired recall rate" Compl. ¶52-constitutes the claimed "comparison."
• Intrinsic Evidence for Interpretation:
  • Evidence for a Broader Interpretation: The claim requires termination "based upon" a comparison, which could be read to cover any termination decision that relies on information derived from comparing the system's output to the target set's characteristics, such as confirming that a sufficient number of target set documents have been found by the system.
  • Evidence for a Narrower Interpretation: The patent describes a specific "target technique" where the process stops "when a sufficient number of documents in the target set have been classified as relevant by the independent search strategy" '374 Patent, col. 4:37-40 This could support a narrower construction requiring a direct check of how many target set documents have been found, rather than a more general progress monitoring against a recall goal.

VI. Other Allegations

Indirect Infringement

• The complaint alleges inducement of infringement under 35 U.S.C. § 271(b) Compl. ¶56 The basis for this allegation is Defendant's publication of "product documentation, product descriptions, blog posts, white papers, videos, and other training materials instructing customers and end users to use the accused functionality" in a manner that allegedly infringes the '374 Patent Compl. ¶58

Willful Infringement

• The complaint alleges willful infringement based on Defendant's alleged knowledge of the '374 Patent since at least March 26, 2026, the date of a notice letter sent by Plaintiff Compl. ¶55 It is alleged that despite this notice, Defendant continued its infringing conduct, demonstrating at least reckless disregard of Plaintiff's patent rights Compl. ¶59 Compl. ¶60

VII. Analyst's Conclusion: Key Questions for the Case

This dispute will likely center on two key questions of claim scope and technical operation:

1. A core issue will be one of definitional scope: Can a "seed set" of documents, provided by a user to initiate and train a continuous active learning algorithm, be legally construed as the claimed "target set of documents," which the patent describes as a formal benchmark used to conduct a specific termination test with a probabilistic guarantee of recall?
2. A key evidentiary question will be one of technical mechanism: Does the accused platform's functionality for stopping a review when a user-defined "desired recall rate is reached" perform the specific act of "terminat[ing]... based upon a comparison between the results of the... search strategy and a characteristic of the target set," or is there a fundamental mismatch between a general progress-monitoring feature and the specific termination condition recited in the claim?