Patent drafting is a complex process that requires a meticulous approach to ensure that inventions are accurately described, claimed, and protected under the law.
In 2024, the advent of advanced AI-powered tools is revolutionizing how patent applications are drafted, offering new efficiencies while maintaining the precision necessary for legal and technical compliance.
This guide explores how AI can be harnessed in the patent drafting process, from understanding the essential elements of a patent application to leveraging AI tools for generating detailed descriptions, claims, and drawings. Whether you are a seasoned patent attorney or new to the field, this guide provides valuable insights into how AI can streamline your workflow, enhance the quality of your patent drafts, and ultimately help secure robust patent protections for innovative technologies.
Elements of a Patent Application: AI Solutions Providing Properly Formatted Outputs
In order for an inventor to be granted patent rights, the inventor must first file a patent application with the agency responsible for registering patent rights in a particular country or region. A patent application is a legal document that claims and fully describes an invention in technical language. While many artificial intelligence applications can produce technical documents, patent applications include a host of formalities that require specialized AI solutions to include those formalities in patent drafting outputs.
A basic list of formalities required in most patent application filings includes:
- Title: The name of the invention to be claimed and described by the patent application.
- Abstract: A short paragraph providing a summary of the invention that is more fully described in the specification.
- Patent Claims: A numbered list of sentences marking out the legal boundaries of the invention that the patentee can enforce against alleged infringers.
- Related Applications: A list of other patent applications filed by an inventor, including divisional or continuation patents, claiming subject matter overlapping with the present patent application. This list also includes foreign patent filings related to the invention.
- Prior Art Citations: A list of patents and other publications discovered by the patent applicant that must be disclosed because they present a material impact to the patentability of the claimed invention.
- Background: A brief portion of the patent’s specification that establishes the technical field of the invention, the current state of the technical art and any drawbacks reducing the utility of that art.
- Brief Description of Drawings: A list providing short descriptions of the subject matter depicted in the technical drawings or diagrams attached to the patent application.
- Detailed Description: A thorough explanation of the inventive subject matter being protected by the patent application’s claims that shows possession of the invention. This section must enable a person of ordinary skill in the related field of art to practice the invention being claimed.
- Drawings: Illustrations or diagrams that either depict the structural qualities of a claimed invention or explain the workflow of a claimed inventive method.
The public disclosure that comes with filing a patent application is the trade-off made by the inventor to obtain the right to prevent others from practicing the invention. As a result, confidentiality is key when drafting a patent application and preparing it for filing. AI-powered patent drafting solutions that are not certified according to industry-leading cybersecurity protocols are not viable options for this reason.
A patent attorney has several tasks to perform prior to drafting a patent application. In order to better understand the invention, a legal professional conducts interviews with one or more inventors and develops documentation providing a full disclosure of the invention to be claimed. Artificial intelligence platforms with security standards that meet the legal industry’s strict confidentiality requirements can use this material as a prompt for generating output without risking a publication that could impact patentability.
The Patent AI Copilot software solution offered by DeepIP has been tested and audited over many months to obtain data security certifications critical to confidential patent drafting. Our platform conforms with ISO/EIC 27001, a standard jointly developed by the International Organization for Standardization and the International Electrotechnical Commission that ensures best practices in managing data security risks. Our AI patent drafting tool also complies with SOC 2, a voluntary standard developed by the American Institute of CPAs that applies trust services criteria to the management of customer data.
Obtaining an Invention Disclosure: Developing Material to Prompt Patent Drafting Software
For most patent attorneys, the first step in the process of drafting a patent application involves collecting information from the inventors about the invention to be claimed. Invention disclosures are studied extensively by patent practitioners in order to understand the nature of the technical advance that can be protected by patent rights.
While many companies utilize standard invention disclosure forms to create business records of an invention, in-person meetings with inventor teams are critical to the patent drafting process. Attorneys should ask questions that help them clearly grasp the novel and nonobvious aspects of an invention, and meeting with inventors in person usually imparts that information to attorneys in ways that improve recall while drafting patent claims or specifications.
In order to collect an invention disclosure that serves effectively as a prompt for patent drafting software solutions, in-house or outside counsel should set a clear agenda for inventor meetings. Productive interviews result in a clear definition of the relevant industry and the current state of competition in that sector. Along with strong communication skills, the ability to interact with different personality styles serves a patent attorney very well in conducting effective interviews that produce ample information for drafting a patent application.
Along with a thorough description of the invention, patent attorneys will want to obtain several other pieces of information that are critical whether or not an AI-powered patent drafting tool is being utilized:
- Prior Art: Previously filed patent applications and scientific publications impact the novelty and nonobviousness of any invention, and inventors are often aware of at least some prior art bearing on patentability that practitioners have a duty to disclose during prosecution.
- Public Disclosures/Commercial Activity: Publications of an invention’s technical advance or offers to sell the invention trigger important legal deadlines that could bar the issue of patent rights if a patent application is not timely filed.
- Inventor Identities: Every member of an inventor team making a significant contribution to the technical advance being claimed must be identified in the patent application. Otherwise, costly legal proceedings may end up being brought by those individuals and threaten the legal enforceability of patent claims.
- Public Funding Sources: Research and development operations that are funded by government entities must be disclosed to a patent granting agency as such funding tends to give governments certain rights in the invention being claimed.
Invention disclosure processes tend to involve some assessment of the commercial viability of the invention, especially in large corporate settings. While that information is important in determining whether to advance a particular invention toward consumer markets, patent attorneys must be careful not to include commercial viability in the patent specification as it has no bearing on patentability. A well-designed artificial intelligence patent drafting solution will help ensure that patent application outputs are properly focused on novelty and nonobviousness rather than potential market success.
Patent practitioners who utilize DeepIP’s artificial intelligence solution for patent drafting can realize time savings from 20% to 40% of the hours spent on writing claims and specifications per patent application. Whereas many patent attorneys spend a great deal of their work week just to understand the invention to be claimed, our AI-powered drafting tool gives attorneys the ability to jump right into drafting after obtaining an invention disclosure. After using the disclosure as an initial prompt, additional queries presented through DeepIP’s embedded word processor tool can provide variations on invention title, patent claims or detailed descriptions to secure broad patent protections.
How to Approach Patent Drafting: Should AI Iterations Start From Claims or Drawings First?
It can be difficult to know where to start a major project like drafting a patent application, but in most instances, patent attorneys have one of two ways to begin. Patent drafting software that utilizes generative AI techniques can refine a patent application over multiple iterations, but the choice of where to start begins with the patent attorney.
Experienced patent attorneys have suggested that there are two sections of a patent application that provide the easiest place for an attorney to begin drafting:
- Beginning with Patent Claims: While this approach may require more upfront study of an invention disclosure to fully understand the subject matter, starting with the claims first creates patent drafting efficiencies for attorneys. Writing a detailed description from complete patent claims tends to result in a concise and precise specification. This approach is more difficult when drafting patent applications for highly complex inventions or those where the full inventive scope of the subject matter may still be unknown.
- Beginning with Patent Drawings: A patent drafting approach that begins with the drawings first can benefit patent attorneys by helping them grasp the full scope of the invention earlier in the drafting process. Revisions can be more easily incorporated than with the claims-first approach and it also helps patent attorneys identify inventive subject matter that leads to drafting more complete claims later in the process. This approach sometimes results in unnecessary extra drawings and inefficiencies in drafting claims last, which can result in inconsistencies with a fully drafted specification.
Once an approach has been selected, a patent attorney must maintain a proper level of detail throughout the drafting process. Too much detail in the claims or specification can narrow a patent application to the point that it doesn’t retain most of the invention’s value. Not enough detail can create enablement or subject matter eligibility issues that risk a patent’s validity at the agency or in court.
Patent attorneys, whether or not they are working with AI-powered patent drafting software, should seek a level of detail that resembles a technical training manual. While the patent application must be drafted to enable a person of ordinary skill in the art to practice the invention, the specification should focus only on the novel aspects of the invention. The specification and drawings must work in tandem so that a skilled artisan can use them to reconstruct an invention as if they were reading an instruction manual.
Longer patent claims or specification do not necessarily insulate a patent application from validity issues. If a lengthy method claim narrows a claimed invention by reciting well-known and routine steps, a court or agency may find the subject matter ineligible for patenting despite descriptions of novel embodiments in the specification. Conversely, a generative AI platform prompted to output a detailed description of incredible length could cause an agency to question whether the patent applicant is actually in possession of a novel invention.
The patent drafting software solution developed by DeepIP is capable of producing formatted patent applications using either an invention disclosure or patent drawings to output a detailed specification with claims. Our generative AI platform uses labels from diagrams and flowcharts to produce descriptions of novel embodiments, or it can develop patent claims useful for drafting a full specification. Our customers also have access to prompt libraries that allow them to drill down on novelty or focus on addressing the technical problem, whichever best suits their drafting style.
Writing Patent Claims: Reviewing Different AI Iterations to Accurately Claim the Invention
While most patent attorneys come to the industry with educational backgrounds in scientific fields, the use of proper grammar is critical to drafting claims that will be upheld as valid in court. Even an AI-powered patent drafting tool designed to produce outputs from invention disclosure prompts in patent application format, cannot be expected to replicate perfect claim grammar at all times. Patent attorneys can save a great deal of time without sacrificing claim drafting quality if they refine claim language over multiple AI iterations.
A typical patent claim has three separate elements making up the entirety of the claim:
- Preamble: The introductory phrase of a patent claim. While the preamble is typically non-limiting, it does define the field of the invention being claimed.
- Transitional Phrase: A short phrase following the preamble that establishes the breadth of the patent claim. Claims typically use either the broader transitional phrase “comprising” or the narrower phrase “consisting of.”
- Limitations: A list of elements or characteristics that defines the entirety of the invention as it differs from the existing prior art.
Even when each of these elements are present, patent attorneys must pay close attention to the language used to construct the claim. Patent claims must be written as a single sentence, which often requires the use of semicolons following each individual limitation. Patent drafting software that utilizes artificial intelligence techniques can recommend several forms of preamble or transitional phrases, giving patent attorneys options during claim drafting that they may not have considered.
Patent attorneys must also be careful to use proper antecedents when introducing or reiterating limitations. For example, in a computer system invention, the processor limitation must be introduced with “a” or “one or more” for multiple processors. Once introduced, proper patent drafting requires that later references to that same limitation use “the” or “said” to refer back to the processor. A patent claim for a computer system having multiple processors may also distinguish those as “first” and “second” processors, especially if they carry out distinct aspects of the invention.
A patent must always begin with an independent claim, which includes only the essential limitations that constitute a claimed invention. Each independent claim may have several dependent claims that serve to further refine the contours of the invention being claimed. A dependent claim following an independent claim for a method might clearly define certain steps of the method, for example. Dependent claims can make a patent more resistant to validity challenges and make it more difficult for competitors to design around a claimed invention to achieve the same result.
DeepIP’s artificial intelligence solution can develop a wide range of independent and dependent claims from inventor interview notes and other documentation to aid attorneys in drafting an entire family of patent applications, giving legal professionals the ability to protect their clients’ interests to the greatest extent possible.
Writing the Patent Specification: Generating Detailed Descriptions That Enable Skilled Artisans
Although a patent’s specification technically includes the patent claims, in most cases patent practitioners use this term to refer to the detailed description that discloses embodiments of the invention in ways that support the granted patent claims. When patentees run into legal issues with their specification, it’s usually because an examiner or judge does not believe the written description enables a person of ordinary skill in the art to practice the claimed invention.
Patent attorneys utilizing AI-powered patent drafting tools can show a patent applicant’s possession of the claimed invention to satisfy the enablement requirement. There is no artificial intelligence tool that can provide a perfect specification after a single prompt, but multiple iterations of a patent specification following several prompts can result in a full, exact and thorough description of invention embodiments that informs others how to make and use the invention.
The enablement requirement is crucial to the underlying rationale providing the foundation for patent law: in exchange for a limited period of market exclusivity, a patentee discloses enough information about the invention that it benefits the public domain once the period of exclusivity expires. Differences in case law from one patent-granting jurisdiction to another can impact the amount of technical description required to meet the enablement requirement. For example, the U.S. Patent and Trademark Office issued a notice this January clarifying examination guidelines for enablement following the U.S. Supreme Court’s 2023 decision in Amgen v. Sanofi. Commentators from the U.S. patent industry have argued that Amgen appears to require disclosure of the full scope of embodiments practicable by a genus claim, which has posed an issue for patentees in the chemical and biopharmaceutical arts.
Regardless of the technical field of the claimed invention, patent practitioners can follow several best practices in drafting specifications to show possession of the invention without undue experimentation:
- Use Specific Language When Describing Technical Embodiments: Although many patent attorneys look to use generic terms to capture a wider scope of the claimed invention, specific technical language is required when describing embodiments to prove possession of the invention. Relying on generic terms generated by AI patent drafting platforms can raise subject matter eligibility concerns along with enablement issues.
- Draft Complete and Thorough Descriptions of the Invention: Remembering that a patent application functions like a technical manual, patent practitioners must draft thorough explanations of how invention embodiments work. Patent drafting software solutions can generate detailed roadmaps for several embodiments of a claimed invention to clearly delineate the boundaries of patent claims.
- Highlight Your Invention’s Novel Structural Qualities: Many patent practitioners find it easy to describe how a claimed invention functions, but focusing on the structure of the invention and how it differs from the prior art can make patent enforcement effective against manufacturers and distributors, not just end users of the invention.
- Clarify Shortcomings in the Existing Art: The background section of a patent specification affords patent applicants the opportunity to differentiate their invention from the prior art by clearly explaining the deficiencies of the current state of the art. Using AI tools to determine shortcomings addressed by the invention can reduce concerns with undue experimentation by informing a skilled artisan of the advance present in the claimed invention.
Whether a patent practitioner is starting their drafting process from diagrams or patent claims, DeepIP’s AI Patent Copilot software platform can quickly generate properly formatted patent specifications that serve as strong first drafts. Our artificial intelligence platform can be prompted to provide additional support for patent claims in the written description, or take a component-driven approach that underscores the structural uniqueness of the invention to meet enablement requirements.
Interested in experiencing yourself how AI can benefit your Patent Drafting? Feel free to request your trial of the tool here!