Phoenix Solutions Inc v. West Interactive Corp

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Phoenix Solutions, Inc. (Maryland)
  • Defendant: West Interactive Corp. (Delaware)
  • Plaintiff’s Counsel: Trojan Law Offices
• Case Identification: 2:09-cv-08156, C.D. Cal., 11/06/2009
• Venue Allegations: Plaintiff alleges venue is proper in the Central District of California because Defendant conducts business and commits infringing acts within the district, and because Plaintiff's principal place of business is located there.
• Core Dispute: Plaintiff alleges that Defendant’s interactive voice response (IVR) systems, which are provided to corporate clients for customer service, infringe four patents related to distributed speech recognition and natural language query processing.
• Technical Context: The technology at issue involves automated systems that allow a user to speak a question in a natural language to a computer and receive an intelligent, spoken response, primarily for use in telephone-based customer service.
• Key Procedural History: The complaint notes that Plaintiff previously filed suit against The DirecTV Group, Inc., a client of Defendant West, on February 13, 2008. In that litigation, DirecTV allegedly argued that West, as the provider of the accused IVR system, was the responsible party for any infringement.

Case Timeline

Date	Event
1999-11-12	Earliest Priority Date ('640, '846, '854 Patents)
2003-10-14	U.S. Patent No. 6,633,846 Issued
2003-12-16	U.S. Patent No. 6,665,640 Issued
2004-03-02	Earliest Priority Date ('431 Patent)
2007-10-02	U.S. Patent No. 7,277,854 Issued
2008-02-13	Plaintiff files suit against The DirecTV Group, Inc.
2009-06-30	U.S. Patent No. 7,555,431 Issued
2009-11-06	Complaint Filing Date

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

U.S. Patent No. 6,665,640 - “Interactive Speech Based Learning/Training System Formulating Search Queries Based on Natural Language Parsing of Recognized User Queries”

The Invention Explained

• Problem Addressed: The patent’s background section describes the limitations of conventional non-voice-based interfaces for internet applications such as remote learning and e-commerce, noting that many users find keyboard-and-mouse interaction unnatural or difficult, and that text-based search engines are often inefficient for finding information (’640 Patent, col. 1:46-2:12).
• The Patented Solution: The invention is a distributed learning or training system that allows a user to ask a question in natural language and receive a spoken answer from a "virtual agent." (’640 Patent, Abstract). The system converts the user’s speech to text, uses a natural language engine to parse the text and identify key components (such as noun phrases), formulates a structured database query based on that analysis, searches a database of question-answer pairs to find the best match, and returns the corresponding answer to the user. (’640 Patent, Fig. 1; col. 7:4-8:5).
• Technical Importance: This approach sought to make online interactions more efficient and accessible, particularly for users who prefer a voice-based environment over traditional text-based interfaces (’640 Patent, col. 2:6-12).

Key Claims at a Glance

The complaint alleges infringement of at least one claim without specifying which (Compl. ¶25). Independent claim 1 is a representative system claim with the following essential elements:

• An interactive learning system comprising:
• a query file for storing topic query entries;
• an answer file for storing corresponding topic answer entries;
• a speech recognition system for generating recognized speech utterance data from a speech-based query;
• a natural language engine that processes the utterance data using morphological and phrase analysis to form recognized speech sentence data; and
• a query formulation system that converts the sentence data into a search query to identify a matching topic query entry and locate the corresponding answer.

U.S. Patent No. 6,633,846 - “Distributed Realtime Speech Recognition System”

The Invention Explained

• Problem Addressed: The patent identifies challenges in deploying speech recognition over networks like the internet, including the computational limitations of client devices ("thin clients"), the need for speaker-independent accuracy without time-consuming user training, and the difficulty of achieving a real-time, conversational response speed (’846 Patent, col. 3:9-5:14).
• The Patented Solution: The invention proposes a distributed, client-server architecture for speech recognition. A client device performs only minimal, computationally inexpensive processing to extract key acoustic features from a user's speech (e.g., Mel-Frequency Cepstral Coefficients, or MFCCs). This compact set of features is transmitted to a remote server, which performs the more intensive processing required to complete the speech recognition and formulate a response. (’846 Patent, Abstract; Fig. 1; col. 16:55-66).
• Technical Importance: This distributed architecture was designed to enable high-performance, real-time speech recognition for a wide variety of devices, including those with limited processing power, thereby making voice-based applications more scalable and accessible over the internet (’846 Patent, col. 6:1-9).

Key Claims at a Glance

The complaint alleges infringement of at least one claim without specifying which (Compl. ¶30). Independent system claim 11 is representative and includes these essential elements:

• A distributed voice recognition system comprising:
• a sound processing circuit to receive a speech utterance;
• a first signal processing circuit that generates a first set of speech data values that are "insufficient by themselves for permitting recognition of words";
• a transmission circuit for formatting and transmitting the first set of speech data in a "streaming fashion" to a second signal processing circuit; and
• the second signal processing circuit, which generates a second set of speech data values sufficient for a word recognition engine to recognize words in the utterance.

U.S. Patent No. 7,277,854 - “Speech Recognition System Interactive Agent”

• Patent Identification: U.S. Patent No. 7,277,854, “Speech Recognition System Interactive Agent,” issued October 2, 2007.
• Technology Synopsis: This patent describes a speech recognition system featuring an "interactive electronic agent." The agent is designed to facilitate a more human-like dialogue by providing prompts, confirming the substance of a user's query, and delivering responses in a conversational manner, thereby improving the user experience. (’854 Patent, Abstract; col. 2:1-4).
• Asserted Claims: The complaint asserts at least one unspecified claim of the ’854 Patent (Compl. ¶35).
• Accused Features: The complaint alleges that West's IVR systems infringe by utilizing "interactive virtual agents created by West to respond to the caller's inquiries in real-time by providing responses in natural speech" (Compl. ¶14, ¶35).

U.S. Patent No. 7,555,431 - “Method for Processing Speech Using Dynamic Grammars”

• Patent Identification: U.S. Patent No. 7,555,431, “Method for Processing Speech Using Dynamic Grammars,” issued June 30, 2009.
• Technology Synopsis: This invention aims to improve speech recognition accuracy and reduce processing latency by using "dynamic grammars." Instead of relying on a single large, all-encompassing grammar, the system selects and loads a smaller, context-specific grammar based on the user's current environment (e.g., the specific topic of a customer service call). This narrows the universe of possible words, allowing for faster and more accurate recognition. (’431 Patent, Abstract; col. 2:1-15).
• Asserted Claims: The complaint asserts at least one unspecified claim of the ’431 Patent (Compl. ¶40).
• Accused Features: The complaint alleges that West's speech recognition platform infringes by utilizing "multiple grammars, speech recognizers and natural languages to understand and accommodate the variety of" different callers and services (Compl. ¶11, ¶40).

III. The Accused Instrumentality

Product Identification

The accused instrumentalities are Defendant West's "advanced speech recognition systems" and "natural language interactive voice response ('IVR') system" platforms (Compl. ¶8).

Functionality and Market Context

The complaint alleges these systems are designed, built, and customized by West for its corporate clients to handle customer service calls (Compl. ¶8, ¶9). The systems are described as having a "distributed" architecture, with some speech-processing operations performed on a "client computing system (such as telephony platform...)" and others on a "separate server computing system" (Compl. ¶13). Functionally, they are alleged to use "speech recognition and natural language engines" to understand spoken questions and "interactive virtual agents" to provide responses in natural speech (Compl. ¶12, ¶14). The complaint alleges West's services are used by "millions of callers across the United States" and generate "hundreds of millions of dollars in revenue per year" (Compl. ¶8). No probative visual evidence provided in complaint.

IV. Analysis of Infringement Allegations

The complaint does not contain a claim chart or specify any asserted claims, instead alleging infringement of "at least one claim" of each patent. The following tables summarize the infringement theory for representative independent claims based on the complaint's narrative allegations.

U.S. Patent No. 6,665,640 Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
a query file for storing a plurality of topic query entries... an answer file for storing a plurality of topic answer entries	West's IVR systems are allegedly tailored to respond with appropriate answers to specific questions based on data collected and studied from actual calls.	¶16	col. 7:15-18
a speech recognition system for generating recognized speech utterance data from... a speech-based query	West’s IVR systems allegedly use "speech recognition engines to break down a caller’s query into specific words understood by the IVR."	¶12	col. 7:19-21
a natural language engine, which processes said recognized speech utterance data using a... phrase analysis to form recognized speech sentence data	West’s IVR systems allegedly use "natural language engines to understand the meaning of those specific words spoken by callers."	¶12	col. 7:22-26
a query formulation system for converting said recognized speech sentence data into a search query	The complaint alleges West's systems determine the "most likely meaning of the caller's specific question" to provide a relevant response.	¶15	col. 7:27-31

U.S. Patent No. 6,633,846 Infringement Allegations

Claim Element (from Independent Claim 11)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
a first signal processing circuit adapted to generate a first set of speech data values from said speech utterance signals, said first set of speech data values being insufficient by themselves for permitting recognition of words articulated in said speech utterance	The complaint alleges that some speech-processing operations for understanding callers are performed on a "client computing system (such as telephony platform or other hardware)."	¶13	col. 39:6-14
a second signal processing circuit being configured to generate a second set of speech data values based on receiving and processing said speech data values	The complaint alleges that "other speech processing operations are performed on a separate server computing system."	¶13	col. 39:5-10
wherein said first set of speech data values are sent in a streaming fashion over said channel before silence is detected	The complaint does not provide sufficient detail for analysis of this element.		col. 40:1-4

Identified Points of Contention

• Technical Questions: A primary question will be whether the general allegations in the complaint can be substantiated with evidence of the specific technical operations claimed in the patents. For instance, what evidence shows that West's systems perform the specific "insufficient" client-side processing required by the ’846 Patent, or that they perform "query formulation" into a structured search query as required by the ’640 Patent, rather than a more general meaning-matching process?
• Scope Questions: A potential issue for the court may be whether the claims of the patents, which are often described in the context of a "learning/training system" (’640 Patent, claim 1) or an "interactive learning system" (’846 Patent, Abstract), can be construed to cover the commercial telephone customer service IVR systems operated by West.

V. Key Claim Terms for Construction

"query formulation system" (’640 Patent, claim 1)

• Context and Importance: The infringement theory for the ’640 Patent may turn on whether West’s system, which allegedly determines the "meaning" of a query to provide an answer, contains the claimed "query formulation system." Practitioners may focus on whether this term requires the creation of a formal, structured database search query (e.g., in SQL), a step not explicitly alleged in the complaint.
• Intrinsic Evidence for Interpretation:
  • Evidence for a Narrower Interpretation: The specification explicitly illustrates a "Text-to-Query Converter" (184) that generates a "Customized SQL Query" to be executed by a "Database Processor" (186). This suggests the term requires a distinct module that creates a formal, machine-readable database query. (’640 Patent, Fig. 1; col. 7:32-40).
  • Evidence for a Broader Interpretation: The abstract describes the invention more broadly as "formulating search queries based on natural language parsing," which a party could argue does not strictly require an SQL-based implementation but could cover any system that systematically translates parsed natural language into a search operation. (’640 Patent, Abstract).

"insufficient by themselves for permitting recognition of words" (’846 Patent, claim 11)

• Context and Importance: This term defines the core architectural division of labor in the distributed system. Infringement will depend on whether the data transmitted from the client-side component of West's IVR is technically "insufficient" for recognition without further processing by the server.
• Intrinsic Evidence for Interpretation:
  • Evidence for a Narrower Interpretation: The specification provides a specific technical example where the client generates only MFCC vectors, and the server must then compute the separate delta and acceleration coefficients required to complete recognition. This suggests "insufficient" means lacking specific, necessary data components for the recognition algorithm. (’846 Patent, col. 16:1-33).
  • Evidence for a Broader Interpretation: The patent also describes the client-side processing more generally as extracting a "sufficient number of acoustic speech vectors," leaving open the question of what precise number or type of vectors would cross the line from "insufficient" to sufficient for recognition. (’846 Patent, Abstract).

VI. Other Allegations

• Indirect Infringement: The complaint alleges that West operates and/or "directing and causing third parties to operate" its IVR systems, which may support a claim for induced infringement (Compl. ¶25, ¶30, ¶35, ¶40).
• Willful Infringement: Willfulness is alleged for all four patents on the grounds that West is "fully aware of Plaintiff's rights, yet has continued to use the infringing IVR system" (Compl. ¶27, ¶32, ¶37, ¶42). The complaint's reference to prior litigation against West's client DirecTV may be used to argue that West had pre-suit knowledge of the patents (Compl. ¶19).

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

• A central evidentiary question will be one of technical implementation: Given the complaint's reliance on "information and belief," what evidence will be produced to show that the internal architecture and data processing steps of West's commercial IVR systems map onto the specific, multi-step claim limitations of the asserted patents?
• A key legal issue will be one of claim construction: Can the term "query formulation system" be construed to cover a system that determines the "most likely meaning" of a user's speech, or does it require the creation of a formal, structured database query as depicted in the patent's specification?
• A dispositive question for willfulness will be knowledge and timing: Does the prior litigation against West's client, as alleged in the complaint, establish that West had pre-suit knowledge of the patents-in-suit and acted with deliberate or reckless disregard of Plaintiff's patent rights?