US 9263039

US Patent 9263039, titled "Systems and methods for responding to natural language speech utterance," has been assigned to Dialect LLC. The inventors are Philippe Di Cristo, Min Ke, Robert A. Kennewick, and Lynn Elise Armstrong. The patent was filed on September 29, 2014, and issued on February 16, 2016.

The abstract describes systems and methods designed to receive both speech and non-speech natural language input, convert these into textual messages, and then execute the contained questions or commands. The invention utilizes context, prior information, domain knowledge, and user-specific profile data to create a natural interaction environment for users across multiple domains. It also focuses on creating, storing, and using extensive personal profile information to improve the accuracy of determining the context of communications and delivering relevant results.

As of April 26, 2026, the patent's legal status is "Expired - Lifetime," with an anticipated expiration date of August 5, 2025, indicating that the patent has expired. Despite its expiration, the patent family has active litigation, including PTAB cases filed in 2025 and US district court cases filed in Washington (2:26-cv-00810), Virginia (1:23-cv-00581), and Texas (7:25-cv-00060, 2:24-cv-00207). While direct CAFC 2026 dockets specifically for patent 9263039 were not directly identifiable through a broad search of CAFC scheduled cases, the listed district court cases may lead to appellate review. The impact of the patent's expiration on ongoing litigation would require further legal analysis.

Here is a plain-language overview of the independent claims:

Independent Claim 1 (Method Claim):
This claim describes a process for a system to understand and respond to a user's spoken natural language. The method involves several steps:

1. Receiving Utterance: The system first captures a user's natural language speech (e.g., a question or command) through a speech unit.
2. Transcribing to Text: This spoken input is then converted into a text message using advanced speech recognition technology, which employs various linguistic rules or extensive vocabulary models.
3. Parsing Keywords: The system then analyzes this text message to pinpoint key words or phrases.
4. Comparing to Context Grammar: These identified keywords are checked against a set of predefined entries within a "context description grammar," which is linked to specific "agents" (software modules designed for particular tasks or domains).
5. Scoring Matches: The system evaluates how well the keywords match the active grammars within this context description.
6. Determining Best Match: Based on these scores, the system identifies the most relevant match.
7. Selecting an Agent: An appropriate software agent is chosen from a collection of available agents based on this best match.
8. Generating Query/Command: A specific, machine-understandable query or command is then created for the selected agent, formatted precisely to suit the requirements of the information source that the agent will interact with.
9. Sending to Source: This formatted query or command is sent to its designated information source (e.g., a database or another device).
10. Generating Response: Finally, a natural language response is generated and delivered back to the user, based on the information received from the target source.

Independent Claim 15 (System Claim):
This claim outlines the hardware and software components of a system designed to perform the method described in Claim 1. The system comprises:

1. Speech Unit: A component configured to receive natural language speech from a user.
2. Multi-Pass Automatic Speech Recognition Module: A software component responsible for converting the received speech into a text message, using dictation or large vocabulary grammars.
3. Parser: A module configured to analyze the text message and identify key words.
4. Context Description Grammar Module: A storage unit holding the entries against which the keywords are compared.
5. Processor: The central processing unit or component that performs several key functions:
   • It calculates scores for active grammars based on keyword comparison.
   • It determines the best match from these scores.
   • It selects the appropriate agent from a group of agents.
   • It generates a machine-processable query or command tailored to the selected agent and its target information source.
   • It sends this query or command to the target source.
6. Response Generator Module: A component that creates a natural language response for the user based on the results obtained from the target source.