Obviousness

Obviousness Analysis under 35 U.S.C. § 103 for U.S. Patent 8,886,536

An obviousness analysis under 35 U.S.C. § 103 determines whether the claimed invention, at the time it was made, would have been obvious to a person having ordinary skill in the art (PHOSITA). This involves identifying a primary prior art reference, finding secondary references that teach missing elements, and articulating a motivation for a PHOSITA to combine these references with a reasonable expectation of success. The priority date for US 8,886,536 is February 6, 2007.

The background of US 8,886,536 highlights that existing voice user interfaces often required users to memorize specific syntaxes, failed to engage in productive dialogue, and thus presented "missed opportunities for providing valuable and relevant information to users" and for marketers to reach consumers. This context provides a strong motivation for a PHOSITA to combine existing technologies to improve user interaction and advertising effectiveness in voice recognition systems.

Independent Claim 1 & 32: Method and System for Providing Promotional Content Based on a Conversation

Claim 1 describes a computer-implemented method involving receiving a first natural language utterance, responding, then receiving a second utterance relating to the first. It further specifies performing speech recognition, determining domain information for the second utterance based on the first, processing recognized words using competitive domain agents to determine an interpretation, and finally determining and presenting promotional content based on this interpretation. Claim 32 is the system counterpart.

Combination: Pell (US 2005/0288929 A1) in view of Busay (US 7,725,493 B1) and/or Plachta (US 2007/0022008 A1).

• Pell's Contribution: Pell teaches a dialogue management system that engages in mixed-initiative conversations, maintaining context across multiple utterances. It explicitly addresses receiving a first utterance, responding, and then interpreting subsequent utterances (a "second ... utterance relating to the first") within that ongoing dialogue. This covers the conversational flow elements of Claim 1 (receiving first utterance, providing response, receiving second utterance relating to first, and processing recognized words to determine an interpretation).

• Busay's and Plachta's Contribution: Busay discloses a system for providing advertisements in response to voice-based search queries, selecting ads based on query content and presenting them. Plachta further details delivering targeted, voice-based advertisements, selecting ads based on user requests (e.g., for navigation or points of interest), and enabling user interaction with those ads (e.g., a voice command to connect to an advertised business). Both references clearly teach the "determining promotional content based on the interpretation" and "presenting the promotional content to a user" elements.

• Motivation for Combination: A PHOSITA in early 2007, aware of the limitations of existing voice interfaces (as articulated in the '536 patent's background), would be motivated to integrate the robust conversational capabilities of Pell (for understanding user intent across multiple turns) with the voice-based advertising mechanisms of Busay or Plachta. The goal would be to deliver more relevant and highly targeted advertisements by leveraging the deeper contextual understanding gained from an ongoing dialogue, rather than just a single query. Improving the effectiveness of voice advertising through better interpretation of user intent would be a clear business and technical driver.

• Addressing Competitive Domain Agents: While Pell processes utterances in context, it doesn't explicitly detail "competitive domain agents." However, the concept of using domain-specific modules or agents to interpret natural language, and selecting the most probable interpretation from competing possibilities, was known in the art prior to the '536 patent's priority date. The '536 patent itself references earlier VoiceBox patent applications (e.g., US 7,640,160, filed Aug. 5, 2005), which describe domain agents competitively generating context-based interpretations by scoring possible interpretations. Therefore, a PHOSITA integrating a conversational system (Pell) with voice advertising (Busay/Plachta) would naturally employ such known robust interpretation techniques, including using multiple domain agents and selecting the best interpretation through a competitive process, to improve accuracy and relevance in a system designed to handle diverse user requests for advertising purposes.

Therefore, the combination of Pell for conversational context, Busay/Plachta for voice advertising, and the known techniques for competitive domain interpretation in speech processing systems, would render Claims 1 and 32 obvious.

Independent Claim 33: Method for Reinterpreting a User's Statement to Provide Promotional Content

Claim 33 describes a method where, after initial interaction (first and second utterances), a third natural language utterance is received. This third utterance is then used to reinterpret the second utterance, and promotional content is determined based on either the original interpretation or the reinterpretation.

Combination: Pell (US 2005/0288929 A1) in view of Busay (US 7,725,493 B1) or Plachta (US 2007/0022008 A1).

• Pell's Contribution: Pell's dialogue system is designed to adapt its understanding throughout a conversation, modifying its interpretation based on new user inputs. This "adaptive misrecognition" or reinterpretation of prior utterances based on current input is explicitly taught by Pell and further confirmed as known prior art by the '536 patent's own reference to US 7,620,549 (filed Aug. 10, 2005), which focuses on "supporting adaptive misrecognition in conversational speech." This covers the elements related to receiving a third utterance and using it to reinterpret a previous (second) utterance.

• Busay's and Plachta's Contribution: As previously established, Busay and Plachta teach the selection and presentation of promotional content in a voice-based interaction.

• Motivation for Combination: Building upon the motivation for Claim 1 (to deliver more targeted voice ads via conversational context), a PHOSITA would further be motivated to ensure that the promotional content delivered is as accurate and relevant as possible. If a conversational system (Pell) can refine its understanding of a user's intent through reinterpretation based on subsequent dialogue, it would be a natural and obvious step to use this improved understanding to select and present more precise promotional content (from Busay or Plachta). This directly addresses the problem of delivering optimal advertisements in evolving conversational scenarios.

Therefore, the combination of Pell's adaptive conversational capabilities (including reinterpretation) with Busay's or Plachta's voice-based advertisement delivery would render Claim 33 obvious.

Independent Claim 37: Method for Using Promotional Content to Clarify Ambiguous Requests

Claim 37 focuses on situations where a user's request is incomplete or ambiguous. The method involves identifying such a request, determining and presenting promotional content related to it, monitoring the user's interaction with that content, and then interpreting the ambiguous request based on that interaction.

Combination: Pell (US 2005/0288929 A1) in view of Plachta (US 2007/0022008 A1) and/or Busay (US 7,725,493 B1), potentially further informed by Rorex (US 2008/0109453 A1) for conceptual targeting.

• Pell's Contribution: Pell describes a dialogue system designed to handle complex data and ambiguous situations, where it might ask clarifying questions to resolve uncertainty. This encompasses the elements of identifying requests and determining when they are incomplete or ambiguous.

• Plachta's and Busay's Contribution: Both Plachta and Busay disclose determining and presenting promotional content in response to a user's voice input. Plachta, in particular, describes interactive voice advertising where users can respond to ads (e.g., with a voice command), thus implicitly teaching the concept of "monitoring interaction of the user with the promotional content." Rorex further teaches sophisticated ad targeting based on conceptual understanding of user queries and profiles, which could inform the selection of relevant promotional content even for ambiguous requests.

• Motivation for Combination: Facing an incomplete or ambiguous request in a conversational system (Pell), a PHOSITA would be motivated to resolve this ambiguity efficiently and in a user-friendly manner. Instead of simply asking for clarification, a PHOSITA, aiming to enhance user experience and capitalize on advertising opportunities (as per the stated problems in the '536 patent), would combine Pell's ambiguity handling with the interactive voice advertising of Plachta or Busay. By presenting a contextually related advertisement (perhaps using conceptual targeting insights from Rorex) and observing how the user interacts with it (e.g., a follow-up voice command or engagement), the system can infer the user's true intent. The '536 patent explicitly states this motivation: "advertisements may be selected to help in interpreting incomplete and/or ambiguous requests." This innovative use of promotional content as a probe to clarify ambiguous requests, leveraging existing interactive advertising paradigms, would be an obvious application of known technologies.

Therefore, the combination of Pell for handling ambiguous conversational requests, along with Plachta (or Busay) for presenting interactive voice advertisements and monitoring user interaction, would render Claim 37 obvious.