PTAB

IPR2025-01240

Amazon.com Services LLC v. VB Assets LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Processing Natural Language Utterances Including Requests and Selecting and Presenting Purchase Opportunities
  • Brief Description: The ’758 patent discloses technology for purchasing products using voice commands. The system processes natural language utterances to determine context, selects corresponding "purchase opportunities," and presents them to a user, potentially using historical user data to build profiles and refine selections.

3. Grounds for Unpatentability

Ground 1: Obviousness over Aretoulaki and Ramer - Claims 1-5, 9-10, 21-27, 31-32, 43-44 are obvious over Aretoulaki in view of Ramer.

  • Prior Art Relied Upon: Aretoulaki (WO Pub. No. 02/073331) and Ramer (Patent 7,752,209).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued Aretoulaki discloses a "knowledge communications platform" for dialogue-based electronic commercial transactions that performs the core method of independent claims 1 and 23. This platform processes natural language speech input, determines context using an NLP Manager, selects a "result set" of products (the "purchase opportunity"), and delivers it to a user's device. Petitioner contended that to the extent Aretoulaki does not explicitly disclose tracking an interaction pattern that includes completing a transaction, Ramer supplies this teaching. Ramer discloses a wireless search platform that tracks user activities, including online product purchases ("completing a transaction"), to build user profiles and infer preferences.
    • Motivation to Combine: A POSITA would combine these references because they address the same field of voice-based e-commerce. Petitioner asserted a POSITA would be motivated to incorporate Ramer’s detailed user interaction tracking (including completed purchases) into Aretoulaki’s system to improve its ability to interpret ambiguous user inputs. This would enhance the system by using more detailed user history to clarify queries and speed up the search and purchase process.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success in combining the references. The proposed modification involved applying Ramer’s known method of tracking user transaction history to Aretoulaki’s system, which already used a "user model," to yield the predictable result of a more robust e-commerce system that could better infer user preferences.

Ground 2: Obviousness over Kennewick and Hao - Claims 1-2, 4-7, 9-24, 26-29, 31-44 are obvious over Kennewick in view of Hao.

  • Prior Art Relied Upon: Kennewick (Application # 2004/0193420) and Hao (Application # 2008/0189187).

  • Core Argument for this Ground:

    • Prior Art Mapping: Petitioner argued Kennewick discloses an interactive speech interface that processes natural language commands, determines context using a parser and domain-specific "agents," and provides responses, including "interactive offers and promotions" for goods and services (a "purchase opportunity"). Kennewick’s system uses a user profile that includes customer order history. Petitioner asserted that Kennewick leaves open the specific method for selecting purchase opportunities based on the determined context. Hao fills this gap by disclosing an e-commerce method that extracts "semantic meanings" from a user’s shopping request, matches them to available inventory items, and automatically adds them to an electronic shopping cart (the "purchase opportunity") for the user to finalize.
    • Motivation to Combine: A POSITA would combine Kennewick and Hao to provide a complete and improved system. Kennewick provides a robust framework for interpreting speech in various domains but is not explicit on how purchase opportunities are selected within an e-commerce context. Petitioner argued that Hao provides a known, straightforward technique for this selection process. A POSITA would be motivated to integrate Hao’s method to improve Kennewick’s remote ordering functionality, enabling the system to automatically analyze a user's request and present a specific purchase opportunity, thereby minimizing the interaction required to deliver a response.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success because the combination involves integrating a known technique (Hao's method for selecting purchase items) into a similar, compatible system (Kennewick's speech interface) to achieve a predictable improvement in functionality.

  • Additional Grounds: Petitioner asserted additional obviousness challenges, including:

    • Adding Jong (Patent 6,173,250) to the Aretoulaki/Ramer and Kennewick/Hao combinations for its specific teaching of "phonetic dictation" to meet limitations in claims 8 and 30.
    • Adding Kennewick to the Aretoulaki/Ramer combination for its teachings on using an "agent architecture" with modular domain agents to meet limitations in claims 6-7 and 11-20.
    • Adding Aretoulaki to the Kennewick/Hao combination for its teachings on monitoring user interactions to detect and repair misunderstandings to meet limitations in claims 3 and 25.

4. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-44 of the ’758 patent as unpatentable under 35 U.S.C. §103.