Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 112
2. The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
3. Claims 10-15 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 10 recites the limitation “the input prompt” in lines 17-21 without antecedent basis. Claims 11-15 recites “the input prompt” and/or inherits the deficiencies of Claim 10. For purposes of examination the input prompt is interpreted as the user prompt.
4. Claims 16-20 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 16 recites the limitation “the input prompt” without antecedent basis. Claims 17-20 recites “the input prompt” and/or inherits the deficiencies of Claim 16. For purposes of examination the input prompt is interpreted as the user prompt.
Claim Rejections - 35 USC § 103
5. In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
6. Claims 1-8 and 10-20 are rejected under 35 U.S.C. 103 as being unpatentable over Hettige (US 2024/0330580) in view of Vouitsis (US 2026/0105056) and further in view of Schmidt (US 2024/0338554).
Regarding Claim 1:
Hettige discloses a computer-implemented method of generating synthetic data for variable replacements in prompts for prompt execution (Hettige: ¶7 discloses a computer implemented method for generating synthesized training examples using prompt templates and response templates, where prompt values and response values are inserted into placeholders and the model learns slot filling for execution plans), the method comprising, by a computing system comprising one or more computing devices:
processing, with at least one machine-learned generative model, an input prompt including a variable for data replacement and at least one data source reference (Hettige: teaches the input prompt including a variable, ¶152 and Table 1 disclose the prompt variables/placeholders {candidate_actions}, {user_profile} etc. ¶62- 63 discloses the data source reference because the context memory store connects external data to large language models (LLMs) from files, documents, datastores, applications programming interfaces (APIs) and metadata from those assets is appended to the utterance to construct the input prompt);
to generate an output indicative of a context of the input prompt
(Hettige: ¶194-196 teaches processing with a machine learned generative model, an input prompt including a variable for data replacement because it discusses prompt templates including prompt placeholders associated with candidate actions, context information and an utterance, where the placeholders may include argument slots having missing values. Hettige further teaches that the utterance provides include argument slots having missing values. Hettige further teaches that the utterance provides information for filling in the missing values);
to generate an output including a (Hettige: ¶195-200 teaches generating prompts by inserting values into prompt placeholders and response placeholders, linking the generated prompt/response as synthetic datapoints and fine-tuning for slot filling where values are inserted into the argument slots based on context).
Hettige does not explicitly:
processing, with a machine-learned embedding model, the data source, the context, and the data type to generate an output indicative of one or more document chunks from the data source for emulation;
processing, with the at least one machine-learned generative model, a first intermediate prompt including the one or more document chunks from the data source to generate an output indicative of one or more attributes of the one or more document chunks; and
processing, with the at least one machine-learned generative model, a second intermediate prompt including the context, the data type, the one or more document chunks, and the one or more attributes
However, Vouitsis discloses:
processing, with a machine-learned embedding model, the data source, the context, and the data type to generate an output indicative of one or more document chunks from the data source for emulation (Vouitsis: ¶55-57 discloses a data source corpus of documents, stored in a repository and subdivided into chunks, ¶70-74 teaches processing with an embedding model each document/chunk is converted by an embedding model into embeddings/vectors and the query is converted using the same embedding model to find closest chunks);
processing, with the at least one machine-learned generative model, a first intermediate prompt including the one or more document chunks from the data source to generate an output indicative of one or more attributes of the one or more document chunks (Vouitsis: ¶116-118 discloses each chunk is provided to a synthetic generation prompt, such as “write a summary for the given passage: {chunk}” to generate synthetic information related to the chunk, including a summary keywords or questions answerable by the chunk. Including a summary, keywords or questions answerable by the chunk); and
processing, with the at least one machine-learned generative model, a second intermediate prompt including the context, the data type, the one or more document chunks, and the one or more attributes (Vouitsis: ¶140-145 discloses the second prompt using query/context, chunks and synthetic information, the retrieval augmented generation (RAG) system retrieves chunks and corresponding synthetic information, provides the chunks, synthetic information and query to the LLM and generates a response based on them).
Hettige and Vouitsis are analogous because both are directed to LLM based systems that process user prompts using context and external source data. Hettige teaches a digital assistant system that constructs prompts from user utterances, context information, candidate actions and asset metadata, where the assets may include APIs, files, documents and data stores. Hettige further teaches prompt placeholders, arguments slots, response placeholders, slot-filling and synthetic datapoints generated from prompt response pairs. Vouitsis teaches a known RAG improvement for retrieving relevant information from such document sources by storing documents in a repository, subdividing the documents into chunks, generating embeddings for chunks and synthetic information, retrieving relevant chunks and using those chunks synthetic information in LLM generation. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Hettige’s prompt execution and slot filling system to use Vouitsis’ chunk based RAG and libraries to generate responses. Vouitsis provides a known technique for retrieving the most relevant portions of such document data sources. The motivation for doing so is that it would improve Hettige’s prompt execution by grounding the slot replacement in retrieved source document chunks rather than relying on only general prompt context, improving relevance as stated in ¶56 of Vouitsis: “Since one or more of the documents may be large, chunking the set of documents 204 may help the pipeline 208 extract relevant content and therefore improve both the retrieval performed by the information retrieval system 220 and the response generation performed by the generation LLM 218, making them more precise and relevant.”
Hettige and Vouitsis do not explicitly disclose and a data type from a data source.
However, Schmidt teaches a data type from a data source (Schmidt: ¶96-103 discloses a data type from a data source because Schmidt teaches inserting, into a dictionary prompt, source data information including the field name, the data type of the field, frequent values and summary statistics and using the generative model completion to describe what the field contains).
Hettige, Vouitsis and Schmidt are analogous because all three concern use of generative models with prompts and source context data to produce improved machine generated outputs. Hettige teaches prompt placeholders, response placeholders, argument slots, context information and synthetic prompt datapoints for slot filling. Vouitsis teaches retrieving source document chunks using embeddings and generating output from selected chunks and related synthetic information. Schmidt teaches using source data characteristics in generative prompts, including field name, field type, frequent values and summary statistics to generate a data dictionary description. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Hettige and Vouitsis system with Schmidt’s data dictionary because once Hettige’s slot is filled using source derived chunks retrieved according to Vouitsis, a person of ordinary skill in the art would have had reason to determine the type and attributes of the relevant source data so that the synthesized replacement value conforms to the data source. Schmidt motivates this in ¶78: “In general, markdown tables can include text formatted to convey analytical data (e.g., the data represented by or used to construct a plot) in a multi-column table to a generative model in a manner that helps the generative model (e.g., the attention mechanism of the generative model) recognize the relationships between the rows, columns, and values of the table, thereby improving the accuracy and relevance of responses provided by the generative model.”
Regarding Claim 2:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 1, further comprising: replacing the variable for data replacement in the input prompt with the synthesized data sample (Hettige: ¶¶193-196 discloses that prompts are generated by inserting prompt values into the prompt placeholders and responses are generated by inserting response values into the response placeholders, and further discloses that response placeholders include argument slots filled with response values);
processing the input prompt with the at least one machine-learned generative model to generate an output based at least in part on the synthesized data sample (Hettige: ¶63 teaches that candidate agent metadata is appended to the utterance to construct an input prompt for the LLM, ¶66 discloses invoking the LLM with the input prompt to generate an execution plan).
Regarding Claim 3:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 1, further comprising:
providing to the at least one machine-learned generative model the input prompt including the variable for data replacement and the at least one data source reference to the data source to be used for the data replacement (Hettige: ¶63 teaches constructing an input prompt for an LLM by appending candidate agent/action metadata to the user utterance and invoking the LLM by passing the input prompt to the LLM. Hettige ¶156 and Table 2 further teaches prompt templates having variables, ¶45 also discloses data sources such as APIs, files/documents and data stores available to agents for execution of actions. Vouitsis ¶52 teaches the referenced data source as a corpus of documents stored in a document repository);
providing to the at least one machine-learned generative model the first intermediate prompt including the one or more document chunks from the data source (Vouitsis: ¶116-117 teaches providing the first intermediate prompt including the one or more document chunks from the data source to a machine learning model because);
providing to the at least one machine-learned generative model the second intermediate prompt including the context of the input prompt, the data type from the data source to be used for the data replacement, the one or more document chunks from the data source, and the one or more attributes of the one or more documents chunks (Vouitsis: ¶116-117 teaches providing a second intermediate prompt including the context of the input prompt and the one or more document chunks because Vouitsis teaches that a generation LLM receives selected chunks, synthetic information related to the chunks, the query and a generation prompt and generates a response based on the selected chunks and synthetic information. Schmidt: ¶96-103 teaches the data type from the data source and source data attributes in a generative prompt because Schmidt teaches inserting into a dictionary prompt, source data information including the field name, the data type of the field, frequent values and summary statistics and using the generative model completion to describe what the field contains. Schmidt also teaches downstream prompts constructed using context data, data dictionary information, summaries and relationship descriptions).
Hettige and Vouitsis are analogous because both are directed to LLM based systems that process user prompts using context and external source data. Hettige teaches a digital assistant system that constructs prompts from user utterances, context information, candidate actions and asset metadata, where the assets may include APIs, files, documents and data stores. Hettige further teaches prompt placeholders, arguments slots, response placeholders, slot-filling and synthetic datapoints generated from prompt response pairs. Vouitsis teaches a known RAG improvement for retrieving relevant information from such document sources by storing documents in a repository, subdividing the documents into chunks, generating embeddings for chunks and synthetic information, retrieving relevant chunks and using those chunks synthetic information in LLM generation. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Hettige’s prompt execution and slot filling system to use Vouitsis’ chunk based RAG and libraries to generate responses. Vouitsis provides a known technique for retrieving the most relevant portions of such document data sources. The motivation for doing so is that it would improve Hettige’s prompt execution by grounding the slot replacement in retrieved source document chunks rather than relying on only general prompt context, improving relevance as stated in ¶56 of Vouitsis: “Since one or more of the documents may be large, chunking the set of documents 204 may help the pipeline 208 extract relevant content and therefore improve both the retrieval performed by the information retrieval system 220 and the response generation performed by the generation LLM 218, making them more precise and relevant.”
Hettige, Vouitsis and Schmidt are analogous because all three concern use of generative models with prompts and source context data to produce improved machine generated outputs. Hettige teaches prompt placeholders, response placeholders, argument slots, context information and synthetic prompt datapoints for slot filling. Vouitsis teaches retrieving source document chunks using embeddings and generating output from selected chunks and related synthetic information. Schmidt teaches using source data characteristics in generative prompts, including field name, field type, frequent values and summary statistics to generate a data dictionary description. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Hettige and Vouitsis system with Schmidt’s data dictionary because once Hettige’s slot is filled using source derived chunks retrieved according to Vouitsis, a person of ordinary skill in the art would have had reason to determine the type and attributes of the relevant source data so that the synthesized replacement value conforms to the data source. Schmidt motivates this in ¶78: “In general, markdown tables can include text formatted to convey analytical data (e.g., the data represented by or used to construct a plot) in a multi-column table to a generative model in a manner that helps the generative model (e.g., the attention mechanism of the generative model) recognize the relationships between the rows, columns, and values of the table, thereby improving the accuracy and relevance of responses provided by the generative model.”
Regarding Claim 4:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 1, wherein:
the at least one machine-learned generative model includes a first machine-learned generative model (Vouitsis: ¶55 explains that the RAG pipeline includes a query modification LLM but also states that the LLM can be combined such that a single LLM may perform the functions of the two or more of those LLMs, so the first machine learned generative model is the single combined LLM in Vouitsis);
processing, with at least one machine-learned generative model, the input prompt comprises processing the input prompt with the first machine-learned generative model (Hettige: ¶66 discloses the input prompt processing in the primary reference, the LLM is invoked by passing in the input prompt and the LLM generates the execution plan based on that input prompt. Vouitsis ¶55 explains the query modification LLM and generation LLM can be implemented as a single LLM rather than separate models. Therefore, Hettige teaches processing the input prompt with an LLM, Vouitsis teaches implementing the relevant LLM functions with a single LLM);
processing, with the at least one machine-learned generative model, the first intermediate prompt comprises processing the first intermediate prompt with the first machine-learned generative model (Hettige: ¶66 discloses that the input prompt processing in the primary reference, the LLM generates the execution plan based on that input prompt. Vouitsis ¶55 discloses that the query modification LLM and generation LLM can be implemented as a single LLM rather than separate models. SO Hettige teaches processing the input prompt with an LLM and Vouitsis teaches implementing the relevant LLM functions with a single LLM, which corresponds to the claimed first machine learned generative model);
processing, with the at least one machine-learned generative model, the second intermediate prompt comprises processing the second intermediate prompt with the first machine-learned generative model (Vouitsis: ¶55 discloses that the generation LLM may be combined with other LLMs in the pipeline such that a single LLM performs multiple functions, including generation, ¶121 discloses that the same single LLM may perform the synthetic generation function and the generation LLM function).
Hettige and Vouitsis are analogous because both are directed to LLM based systems that process user prompts using context and external source data. Hettige teaches a digital assistant system that constructs prompts from user utterances, context information, candidate actions and asset metadata, where the assets may include APIs, files, documents and data stores. Vouitsis teaches that and LLM and RAG system can be performed by separate or a single model. It would have been obvious to one of ordinary skill in the art before the effective filing date to incorporate using single LLM implementation for processing the input prompt, first intermediate prompt and a second intermediate prompts because doing so applies as known LLM configuration to reduce model separation while preserving the same prompt processing functions and improving precision when source documents are used as disclosed in Vouitsis “a single LLM may perform the functions described as being performed by two or more” in ¶55 and further discloses ¶56 that chunking source documents improves retrieval and response generation, “making them more precise and relevant.”
Regarding Claim 5:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 1, wherein:
the at least one machine-learned generative model includes a first machine-learned generative model and a second machine-learned generative model (Vouitsis: ¶[0055] teaches separate LLMs in the pipeline, query modification LLM 214, re-ranker LLM 216 and generation LLM 218. This supports using more than one machine learning generative model);
processing, with at least one machine-learned generative model, the input prompt comprises processing the input prompt with the first machine-learned generative model (Hettige: ¶66 teaches invoking an LLM by passing in the input prompt and generating an execution plan based on that input prompt. Vouitsis ¶58-59 teaches the first LLM type operation, query modification LLM 214 receives the user query and query modification prompt to generate synthetic information / modified query);
processing, with the at least one machine-learned generative model, the first intermediate prompt comprises processing the first intermediate prompt with the second machine-learned generative model (Vouitsis: ¶121 teaches that the synthetic generation LLM 926 may be combined with the generation LLM 218, meaning the same LLM can perform the synthetic generation function and the generation function);
processing, with the at least one machine-learned generative model, the second intermediate prompt comprises processing the second intermediate prompt with the second machine-learned generative model (Vouitsis: ¶121 teaches that the synthetic generation LLM 926 may be combined with the generation LLM 218, meaning the same LLM can perform the synthetic generation function and the generation function. ¶145 discloses the second intermediate generation prompt, generation LLM 1108 receives the subset of chunks, synthetic information related to the chunks, the query and a generation prompt and generates a response based on the chunks and synthetic information).
Hettige and Vouitsis are analogous because both are directed to LLM based systems that process user prompts using context and external source data. Hettige teaches a digital assistant system that constructs prompts from user utterances, context information, candidate actions and asset metadata, where the assets may include APIs, files, documents and data stores. Vouitsis teaches that and LLM and RAG system can be performed by separate or a single model. It would have been obvious to one of ordinary skill in the art before the effective filing date to incorporate using single LLM implementation for processing the input prompt, first intermediate prompt and a second intermediate prompts because doing so applies as known LLM configuration to reduce model separation while preserving the same prompt processing functions and improving precision when source documents are used as disclosed in Vouitsis “a single LLM may perform the functions described as being performed by two or more” in ¶55 and further discloses ¶56 that chunking source documents improves retrieval and response generation, “making them more precise and relevant.”
Regarding Claim 6:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 1, wherein: the at least one machine-learned generative model includes a large language model (Hettige: ¶67 teaches the claimed large language model. it explains that LLM 216 has a deep generative model architecture and generates the execution plan using autoregressive language modeling within a transformer architecture).
Regarding Claim 7:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 6, wherein: the large language model is a multimodal large language model (Schmidt: ¶219 teaches the multimodal generative model, it says that instead of transforming visual quantitative data into text and providing it to a text based generative model, “a multimodal generative model” may be used, where the model is capable of processing text data and other data types, such as image data or video data).
Hettige, Vouitsis and Schmidt are analogous because all three concern use of generative models with prompts and source context data to produce improved machine generated outputs. Hettige teaches prompt placeholders, response placeholders, argument slots, context information and synthetic prompt datapoints for slot filling. Vouitsis teaches retrieving source document chunks using embeddings and generating output from selected chunks and related synthetic information. Schmidt teaches using source data characteristics in generative prompts, including field name, field type, frequent values and summary statistics to generate a data dictionary description. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Hettige and Vouitsis system with Schmidt’s multimodal input. The motivation for doing so is “ visual quantitative data are included in a prompt provided to the multimodal generative model without transforming the visual quantitative data into text.” Therefore showing the benefit of including other forms of data to provide directly to the prompt rather than first converting that data into text.
Regarding Claim 8:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 1, further comprising:
providing a graphical user interface including a first user interface element configured to receive the input prompt and a second user interface element configured to receive the at least one data source reference to the data source (Hettige: ¶48 discloses a user interface because it says a user may provide user inputs to the digital assistant and receive responses via a user interface element such as a chat window).
Regarding Claim 10:
Claim 10 has been analyzed with regard to claims 1 (see rejection above) and is rejected for the same reasons of obviousness as used above.
It is noted that Hettige discloses a computing system capable of performing the method of claim 10 at least at ¶249 and Fig. 10. Additionally, the user prompt is equivalent to the input prompt.
Regarding Claim 11:
Claim 11 has been analyzed with regard to claims 2 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 12:
Claim 12 has been analyzed with regard to claims 3 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 13:
Claim 13 has been analyzed with regard to claims 4 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 14:
Claim 14 has been analyzed with regard to claims 5 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 15:
Claim 15 has been analyzed with regard to claims 8 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 16:
Claim 16 has been analyzed with regard to claims 1 (see rejection above) and is rejected for the same reasons of obviousness as used above.
It is noted that Hettige discloses a non transitory computer readable media that collectively stores instructions that are capable of performing the method of claim 10 at least at ¶249 and Fig. 10. Additionally, the user prompt is equivalent to the input prompt.
Regarding Claim 17:
Claim 17 has been analyzed with regard to claims 2 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 18:
Claim 18 has been analyzed with regard to claims 3 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 19:
Claim 19 has been analyzed with regard to claims 6 (see rejection above) and is rejected for the same reasons of obviousness as used above.
Regarding Claim 20:
Claim 20 has been analyzed with regard to claims 8 (see rejection above) and is rejected for the same reasons of obviousness as used above.
7. Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Hettige in view of Vouitsis, further in view of Schmidt and further in view of Hariri (US 2024/0330580).
Regarding Claim 9:
The proposed combination of Hettige, Vouitsis and Schmidt further discloses the computer-implemented method of claim 8, except further comprising:
replacing the variable with the synthesized data sample (Hariri: ¶84 discloses user approval and insertion input);
processing the input prompt with the at least one machine-learned generative model to generate an output based at least in part on the synthesized data sample (Hettige: ¶50-52 discloses constructing an input prompt for LLMs and generating an execution plan).
Hettige does not explicitly disclose:
providing, via the graphical user interface, an indication of the synthesized data sample;
receiving, via the graphical user interface, an input to replace the variable with the synthesized data sample;
However, Hariri teaches
providing, via the graphical user interface, an indication of the synthesized data sample (Hariri: ¶83 teaches displaying the generated synthesized output in the GUI before insertion);
receiving, via the graphical user interface, an input to replace the variable with the synthesized data sample (Hariri: ¶84 discloses user approval and insertion input).
Hettige and Hariri are combinable because both concern user interaction with machined learned generative models through prompts and user interfaces. Hettige teaches processing input prompts with LLMs for execution planning and slot filling. Hariri teaches the specific GUI approval insertion not expressly detailed by Hettige. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to disclose Hettige’s prompt replacement system to include Hariri’s GUI preview and insert workflow because Hariri teaches that integrating generative model output directly with user working environments “can substantially reduce the time required by users. In turn, this eliminates the expenditure of substantial quantities of computer resources that a user would otherwise use” in ¶102.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to IAN SCOTT MCLEAN whose telephone number is (703)756-4599. The examiner can normally be reached "Monday - Friday 8:00-5:00 EST, off Every 2nd Friday".
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Hai Phan can be reached at (571) 272-6338. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/IAN SCOTT MCLEAN/Examiner, Art Unit 2654
/HAI PHAN/Supervisory Patent Examiner, Art Unit 2654
Prosecution Insights