RFI Attachment 5 - Questions and Answers as of 22 July 2025.xlsx
XLSX spreadsheet 27 KB Posted
- Solicitation number
- Not on record
About this file
This document is a Request for Information (RFI) Questions and Answers spreadsheet for the United States Patent and Trademark Office (USPTO) seeking innovative solutions to transform patent application and publication processes. The RFI focuses on improving intellectual property data ingestion, processing, and publication technologies, with a goal of moving 100% of patent publications in-house. Key technical areas of interest include complex work unit (CWU) processing, cloud-based solutions, AI/ML capabilities for data conversion and classification, and integration with existing USPTO systems. The RFI seeks vendor experiences and skills in areas such as OCR, image and XML processing, front-end and back-end technologies, and handling diverse patent application formats.
The document indicates the RFI response deadline is July 30, 2025, at 5:00 PM Eastern Time, with a 20-page response limit excluding cover sheets and tables of contents. The USPTO is open to various technical approaches, including human intelligence, AI, and hybrid models for data validation and processing. Current patent application volumes are approximately 325,000 e-Grants and 450,000 Pre-Grant Publications annually, with 97% of applications filed electronically. The agency has not yet determined contract type, set-aside status, or specific implementation timelines, emphasizing this is a market research exercise to understand potential technological solutions for patent processing transformation.
View the file
Other files
On GovTribe
Work with this file
- Download the original file
- Contacts named in this file
- Similar government files
- Ask GovTribe AI about this file
Text version
Q&A
| Question # | Reference | Question Category (Contract or Technical) | Question | Answer |
| 1 | III | Technical | Approximately 40% of applications include Complex Work Units (tables, sequences, chemistry, equations). Can you share examples of CWUs how they look like that most frequently trigger manual intervention? | There is wide range of complexity and variety in applicant's submission of CWUs. CWU solution may require manual correction. As it relates to providing examples, each instance varies so greatly that specific examples would not be representative of the number, or types of CWUs that require manual intervention. |
| 2 | III | Technical | Certificates of Correction often amend multiple sections. Do you anticipate handling these via the same pipeline as Grants, or as a separate CWU stream? | Anticipated to handle these via the same pipeline as Grants. |
| 3 | V | Technical | What level of XML schema validation (DTD vs. XSD vs. Schematron) is currently used in front-end processing, and where are your most significant pain points in QA tooling? | Output is validated against DTD; however, also a 'business rule' content validator application. |
| 4 | Attachment 3 – Grant Workflow | Technical | For the Initial Data Capture (IDC) → Final Data Capture (FDC) transition, what tools or reports (e.g., DREAD) would you like integrated into our UI for examiner review? | Anything that flags deficiencies or changes that need to be reconciled. |
| 5 | V | Technical | Can you outline the core languages, frameworks, and libraries used in your front-end and back-end processing pipelines? | Front-end Vanilla Javascript (jQuery + custom patterns); Back-End: Primarily Java on Wildfly. |
| 6 | V | Technical | Which database and storage technologies (e.g., relational, NoSQL, object stores) are employed for holding XML, images, and metadata? | Oracle and file storage (Network File System (NFS) on premises; EFS/S3 in AWS). |
| 7 | V | Technical | Which manual-review tooling features are most critical today (e.g., side-by-side image ↔ XML compare, annotations, exception-queue management)? | Side by side. |
| 8 | V | Technical | Please clarify what “cloud-based” entails for this solution. Specifically, do you require a public cloud (e.g., AWS, Azure, GCP) with a particular FedRAMP authorization, a private cloud within the USPTO network, or a hybrid model? | While USPTO Fedramp Cloud implementation occurs most commonly using AWS, the Agency has no preference and is able to integrate with a variety of solutions. |
| 9 | V | Technical | Which error metrics (OCR misreads, misclassified objects, missing drawings) are tracked most closely, and what dashboard KPIs would you like to see? | All error metrics are tracked. Dashboard KPI could include frequency of error, criticality of error, error introduced at what step, trendline, recording continuous improvement, among other standard KPIs. |
| 10 | V | Technical | Do you have any preferred AI tools, platforms, or solution approaches that you would like us to consider for the Allowance Initial Data Capture (IDC) and Final Data Capture (FDC) steps? | No; we seek information on experiences and skills to address this challenge. |
| 11 | II - V | Technical | What overall implementation timeline and key milestones (e.g., design, development, testing, pilot, full deployment) do you anticipate under this RFI? | The goal of the RFI is to better understand what the marketplace has to offer in terms of overall capabilities, processes and innovation for the betterment of Patent processing. Overall implementation timeline and key milestones would be addressed after the RFI. |
| 12 | II - V | Technical | Are there planned future phases—such as AI-based classification of objections or AI-examiner assistants—that we should architect our solution to support from day one? | The USPTO seeks to understand what experiences and skills you have to address the problem. The agency position is that AI can be a force multiplier. |
| 13 | Attachment 1 - USPTO and PDDM Workflow | Technical | Can you confirm which specific tasks you expect the contractor to perform during the “Assemble Paper Grant” step (Stage 8), such as document formatting, quality control inspections, binding/packet assembly, and mail preparation? | Ensuring the printed patent matches with the mailing label, quality control inspections, binding/packet assembly, and mail preparation. |
| 14 | IV | Contract | Does the required cover page count toward the 20 page response limit? | No. A cover page will not count towards the 20 page limit. |
| 15 | IV | Contract | May the vendor submit a cover letter in addition to the cover page? If so, will the cover letter count toward the 20 page limit? | Yes. A cover letter will not count towards the 20 page limit. |
| 16 | IV | Contract | May the vendor include a table of contents, figure list, and table list? If so, will they count toward the 20 page limit? | Yes. A table of contents, figure list, and table list will not count towards the 20 page limit. |
| 17 | III | Technical | Is the USPTO seeking a fully developed, vendor-owned solution, or is it looking for a solution approach that can be custom-built to meet the objectives outlined in this RFI? Based on the language in Constraint 1, it appears the USPTO intends to custom develop these capabilities, retain ownership of the tool, and operate it independently. Please confirm. | That is correct, the USPTO will retain ownership of the tool and operate indepedently. |
| 18 | III | Contract | If a vendor offers pre-built capabilities that can accelerate development, how will they be compensated for those components in the event the resulting contract is awarded as time-and-materials? | The USPTO is still conducting market research and details such as contract type have not been determined. If the USPTO proceeds with a procurement, this question will be taken into consideration. |
| 19 | Contract | What is USPTO’s projected timeframe for release of an RFP? | The USPTO is still conducting market research. Therefore, a projected solicitation timeline is still to be determined. | |
| 20 | Contract | Assuming a contract is awarded, what is a desired timeframe for start-up / implementation? | The USPTO is still conducting market research and other acquisition planning activities. Therefore, start-up and implementation timelines are to be determined. | |
| 21 | III | Technical | What capabilities or software products does USPTO use today to facilitate data ingestion and manipulation? | Patent Center is the outward facing tool that applicants file applicatons and follow on papers with the USPTO. The data is loaded in to databases and transformed throughout the process to meet the desired image and xml text standards. |
| 22 | III | Technical | What capabilities or software products does USPTO use today to parse complex work units, including images of data tables, mathematical equations, and chemical formulas? | The USPTO has been initalling addressing classifying the CWU on receipt; USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. |
| 23 | III | Technical | What is the level or success and/or accuracy rate of the current tools USPTO uses to parse complex work units? | It is too early in the implementation to know. USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. |
| 24 | III | Technical | Is USPTO looking to transform/automate both the Publication/Dissemination process as well as the process to provide certified copies of patents? | The USPTO is open to the idea of transforming/automating the process as mentioned. |
| 25 | Contract | Is the government planning to make any potential future solicitation that arises from this RFI a small business set aside? Or will any future solicitation be free and open? | The USPTO is still conducting market research and a set-aside determination has not been made at this time. If the USPTO proceeds with a procurement, an appropriate set aside determination will be made at a later date. | |
| 26 | III | Technical | Approximately how many Patent Applications per year has USPTO been receiving over the last few years? (We understand that patent application volume will vary, but the historical average load is a good place to start to assure that the correct workflow tools will be used within the USPTO cloud). | Information available at Patents Dashboard |
https://www.uspto.gov/dashboard/patents/production-unexamined-filing.html
| 27 | Attachment 2 - PG-Pub Workflow | Technical | Attachment 2 (“RFI Attachment 2 - PG-Pub Workflow.pdf”) seems to indicate a 9-week process lead time but does not indicate volume information. What is the volume of work that is currently sent to the Publication Vendor, and which would be brought in-house via this RFI? | Approximately 7,500 per week. USPTO seeks to potentially produce the requirements internally. However, each vendor is encouraged to provide information that they believe will produce the best value to the government. | |
| 28 | III | Technical | Does your current Publication Vendor currently produce both digital and paper based products? | The Publication Vendor produces a digital product and the printed patent is handled by a different vendor that is a small business. The RFI seeks input on best ways to accomplish both digital and print requirements. | |
| 29 | III | Technical | Is the current Publication Vendor producing Color or Black and White for the paper based products? | The Publication Vendor produces a digital product; the printing of the patent is handled by a different vendor. The RFI seeks input on best ways to accomplish both digital and print requirements in both B&W and color. | |
| 30 | III | Technical | What type of binding(s) is currently being provided by the Publication Vendor? | Glue based book binding. | |
| 31 | III | Technical | Does the current Publication Vendor inventory the paper based products or are they print on demand (POD)? | Patent covers are printed and inventoried. Inside pages are printed on a weekly basis for the patents issued. In addition, there could be 0 to 10 patents printed on demand per week. | |
| 32 | III | Contract | What is the name of the current Publication Vendor? | The Publication Vendor is Reed Technology and Information Services, LLC. | |
| 33 | III | Technical | Are multiple language products in both Digital , Paper required? | Yes. | |
| 34 | III | Technical | Are ADA compliant products i.e., braille, big type, and screen reader a requirement? | Not required at this time. Open to consideration. | |
| 35 | III | Technical | How many of your applications and updates/addendums are received paper vs. electronic | ||
| A) Are applications and their artifacts often a hybrid of paper and electronic components? | 97% applications and updates/follow on documents are filed electronically. Approx 3% are filed on paper. | ||||
| 36 | III | Technical | Are you looking for replacements for just the portions of the workflow that are currently vendor-supplied? Or just one portion of the vendor-supplied components? | USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 37 | III | Technical | Are you looking for CWUs to be broken out of the documents and classified/stored/tagged on their own? Or are you looking for metadata tagging of components within the document itself? | USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 38 | III | Technical | Are you looking for higher-level AI (insight) tagging of CWUs such that it increases the ability of search engines to associate items that might otherwise seem disparate? e.g.: |
A)Simple tagging - “math formula” B) Complex/AI-assisted Tagging- “type: math formula” “subtype: reduction equation” “subtype:friction elimination” “industry:manufacturing” “research:materials science” “geographical usage:Southeast Asia”
| Please refer to the USPTO Open Data Portal (https://data.uspto.gov/home) for examples of Pre Grant Publications and Patent Grants for the level of mark up in the standards today. Futher, the USPTO seeks knowledge and experience in the betterment of Patent processing and increased innovation. | ||||
| 39 | III | Technical | Do you favor a human-intelligence, AI, or hybrid approach to data validation? | USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. |
| 39 | Contract | Does the 20-page limit include the cover sheet, table of contents, graphics, or appendices? | A cover sheet and table of contents will not count towards the 20-page limit. Graphics and appendices will count towards the 20-page limit. | |
| 40 | Contract | When will the question responses be released? Can the USPTO please allow atleast 10 business days between q/a addendum release & submission deadline to allow vendors sufficient time to incorporate all necessary information in their response. | Questions will be released as an amendment to the SAM.gov posting for this RFI. The submission deadline was extended to July 30, 2025 at 5:00pm Eastern Time. | |
| 41 | III | Technical | Can USPTO share hourly or daily surges for PGPub (~450 k/yr) and eGrant (~325 k/yr) workloads so we size compute resources correctly? | Applicant filings vary on a day to day basis. |
| 42 | III | Technical | Must the solution perform OCR and automatic translation of foreign-language content before publication? If yes, which languages and annual volumes apply? | OCR (yes), foreign language translation (no) |
| 43 | III | Technical | Is there any intent to re-convert legacy patents or is the requirement strictly forward-looking? Please provide more information on the scope of reconverting legacy patents. | At this time, we are forward looking. |
| 44 | III | Technical | Can you please provide the following information on your integration requirements as specified in Constraint 2: |
a. List of existing systems that the solution should integrate with.
b. For each of the above-listed existing systems, please describe the expected workflow and scope of integration.
| c. For each of the above listed existing system integration, please sepcify what business challenge USPTO wishes to address through this integration? | At this point, and in terms of solutions and skills, we are seeking information about a broad issue to understand what the marketplace has to offer. This question is more of an implementation question that would be addressed after the RFI. Please provide your experience intergrating with systems and workflows. | ||||
| 45 | III | Technical | What quantitative thresholds (e.g., word-error rate, table-cell accuracy) define acceptance for text/image conversions? | Above 99% accuracy | |
| 46 | III | Technical | Does “solution must be cloud based” (Constraint 3) imply deployment into USPTO-owned CSP accounts, or is a vendor-managed government-cloud enclave acceptable? | USPTO cloud solution only. | |
| 47 | III | Technical | Do the CWUs that require specialized conversion (e.g., mathematical equations, chemical structures) follow any consistent formatting standards (e.g., MathML, ChemDraw, LaTeX), or are they varied and unstructured? | Applicants provide CWUs in a varitey of formats some of which are structure. We currently accept application in DOCX format. We are seeking experience and knowledge in converting these CWU to publication standards. Please look at USPTO Open Data Portal (https://data.uspto.gov/home) for examples of Pre Grant Publications and Patent Grants for the level of mark up in the standards today. | |
| 48 | III | Technical | Is the USPTO looking for automated tools to classify and segment different CWU types, or is manual tagging acceptable as part of the process? | Both approaches are acceptable; please share why one would be more efficient than the other. | |
| 49 | III | Technical | Can the USPTO clarify whether the 25,000 Certificates of Correction per year follow the same format as eGrants or require a different processing pipeline? | Proposed soultion should be following the same pipeline as eGrants. | |
| 50 | III | Technical | Regarding Constraint 5 (AI tools not internet-exposed), can USPTO clarify if models must be pre-trained prior to deployment, or will in-house fine-tuning/training on USPTO data be permitted post-deployment? | USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 51 | III | Technical | For Constraint 6 (handling confidential patent data), is there a requirement for FedRAMP authorization for all cloud components of the solution? | Yes | |
| 52 | III | Technical | Does the USPTO require any specific audit logging, traceability, or explainability features for AI/ML outputs used in conversion and categorization tasks? | At this point we do not have specific requirements. USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 53 | III | Technical | Are there any plans to introduce APIs for patent applicants to submit more structured input formats (e.g., XML, DOCX with style constraints) to simplify downstream conversion? | There are plans to improve our applicants offerings and to improve data exchange. There is no specific direction to share at this time. | |
| 54 | III | Technical | Will there be phased rollouts or pilot opportunities before full deployment across all patent workflows (e.g., starting with eGrants, then PGPubs)? | This is an implementation question; however, USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 55 | III | Technical | Will USPTO provide a ground-truth dataset for initial validation and benchmark testing of the proposed solution during evaluation? | Not at this time; however USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 56 | III | Technical | Does the USPTO plan to conduct formal user acceptance testing (UAT) or sandbox validation with live data prior to production deployment? If so, what will be the evaluation criteria? | This is an implementation question; however, USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 57 | III | Technical | For solutions requiring machine learning, will USPTO permit access to historical labeled data for training and testing purposes under controlled conditions? | Not at this time; however USPTO seeks the experience and skills in the RFI to understand the best way to address this challenge. | |
| 58 | Contract | Is there a defined budget or funding range allocated for this initiative? If so, can the USPTO share whether there are any cost ceilings, pricing constraints, or preferred contract structures ? | The USPTO is still conducting market research and a funding range, cost ceiling, price constraints, and contract structure have not been determined. |
File details come from the government source that posted it. Updated .