.png?width=500&height=250&name=%E7%9F%A5%E7%9A%84%E8%B2%A1%E7%94%A3%E4%BA%8B%E5%8B%99%E6%89%80%E3%82%A8%E3%83%9C%E3%83%AA%E3%82%AF%E3%82%B9%20(500%20%C3%97%20250%20px).png)
In recent years, with the evolution of large language models (LLMs) such as ChatGPT, the...
In today’s world, where driving digital transformation (DX) is considered an urgent priority, there has been a rapid increase in the number of companies leveraging “generative AI” tools such as ChatGPT, Claude, and Gemini to develop custom business efficiency tools and internal dashboards.
Examples include “a dashboard that feeds past internal sales reports into AI and instantly displays answers when employees ask questions” and “a system that automatically classifies and summarizes customer inquiries and presents the optimal solution on the representative’s screen.”
When developing such proprietary systems, many business leaders and DX managers ask, “Can we obtain a patent for internal operational efficiency systems or dashboards created using generative AI?”
Conclusion: While it is difficult to obtain a patent for a system that “simply connects to a generative AI API,” it is entirely possible to obtain a patent if certain conditions—such as unique innovations—are met.
In this article, a patent attorney with expertise in the AI and IT fields provides a thorough explanation of the requirements for obtaining a patent for operational efficiency tools and internal dashboards that utilize generative AI, cases where patents are likely to be granted and those where they are not, and the powerful business benefits of obtaining a patent—even for internal tools.
Key Points of This Article
- Business efficiency tools and internal dashboards utilizing generative AI can be patented if they meet certain conditions
- “Simply using an API” is highly likely to be rejected for lack of inventive step
- Proprietary data pre-processing and post-processing, automatic prompt generation, and dynamic UI control are key to patentability
- Even internal tools offer benefits for SaaS sales, business protection, and fundraising
- The inventor must always be a “human.” AI cannot be an inventor (under Japanese law)
- Publishing press releases or tech blog posts before filing a patent application carries a high risk of loss of novelty
Table of Contents
Chapter 1: Can Tools Created with Generative AI Even Be Patented?
Business efficiency tools and internal dashboards that utilize generative AI are eligible for patent protection under patent law as “software-related inventions” or “AI-related inventions.”
A patent right is the exclusive right to practice a new technical idea (invention). Today, programs, software, and business models that utilize them are widely recognized as patentable. The Japan Patent Office is also focusing on protecting AI-related technologies, and the number of patent grants is increasing year by year.
A common misconception here is the question: “Can program code written by AI be patented?”Under patent law, “who (or what) wrote” the source code of a program is not a direct obstacle to obtaining a patent. This is because the patent system protects not the string of characters in the source code itself, but rather the “technical idea (invention)” realized by that program.
Important Note: Generative AI itself is already a widely available, existing technology. A mere idea—such as “streamlining operations using generative AI”—is not patentable. Furthermore, a “prompt”—an instruction given to AI—is, on its own, merely a human-defined set of instructions and is generally not considered an “invention.”
To obtain a patent, the entire system must meet the “patent requirements” stipulated by patent law. This will be explained in detail in the next chapter.
Chapter 2: Three Essential Requirements for Obtaining Software and AI Patents
For a system utilizing generative AI to be recognized as a patent, it must meet the following three requirements:
1. Qualification as an Invention (Interaction with Hardware)
Under patent law, an “invention” is defined as “a creative technical concept utilizing the laws of nature.” Since software alone is considered merely a computational procedure, to be patentable, the specification must state that “the information processing performed by the software is specifically implemented using hardware resources such as a CPU or memory.”A concrete link to hardware—such as “the server’s CPU retrieves information from a database, inputs it into an AI model to execute processing, and outputs the results to a display”—is required.
2. Novelty (Is it unknown to the public?)
“Novelty” means that the invention has not been made public at the time of the patent application. No matter how groundbreaking a dashboard may be, if its specifications are announced in a press release or its algorithm is published on a tech blog prior to filing, it will, in principle, lose its novelty and become ineligible for a patent. The golden rule is to “file a patent application before disclosing the invention to the public.”
3. Inventive Step (Is it something a person skilled in the art would not readily conceive of?)
The biggest hurdle in AI patents is “inventive step.” This means that the invention must include an improvement that a person skilled in the art could not readily conceive of based on existing technology.
Simply “sending text data to an API and displaying a summary on a dashboard” is considered a design concept that any programmer could conceive of and will be rejected.For inventive step to be recognized, technical features that other companies cannot imitate—such as “the company’s unique data preprocessing and postprocessing methods” or “a mechanism for dynamically generating prompts”—are essential.
Chapter 3: Specific Examples of “Cases That Qualify for a Patent” and “Cases That Do Not”
Specifically, what kind of dashboard can be patented? We’ll explain examples that illustrate the difference between success and failure.
✗ Case That Is Not Patentable: Simple Use of an API
Overview: A chatbot that feeds text data from an internal Q&A system into the ChatGPT API and displays the returned responses directly on a dashboard.
Reason: This is simply using an existing API according to its specifications, with no original technical ingenuity in the pre- or post-processing of the data. It is likely to be judged as “merely automating business operations using general-purpose AI” and is highly likely to be rejected for lack of inventive step.
○ Patentable Case 1: Unique Data Extraction and Automatic Prompt Generation
Overview: The company’s specialized manuals are converted into a vector database, and ambiguous questions from employees are “automatically converted into company-specific technical terms” for searching. The system weights the extracted data based on importance, automatically constructs the optimal prompt, and has the AI generate a response (an advanced application of RAG).
Reason: Because it involves not only the use of RAG (Retrieval-Augmented Generation) technology but also unique innovations in the data processing steps—such as “automatic question conversion” and “weighted prompt construction” (system-side algorithms)—it has the potential to be a strong patent.
○ Patent Case 2: Dynamic Control Linked to UI/UX
Overview: A dashboard for call center operators. This system transcribes and performs sentiment analysis on customer call audio in real time, and dynamically adjusts the “placement of optimal response manuals” and the “color and flashing patterns of alerts” on the dashboard based on the resulting scores.
Reason: Rather than simply displaying the AI’s analysis results as text, the system links them to “dashboard display control (dynamic UI changes).” Since this involves specific hardware (display) control designed to solve a particular problem, it is more likely to be recognized as patentable.
Chapter 4: Three Benefits of Obtaining a Patent Even for Internal Tools
Some may wonder, “Is it worth spending money on a patent attorney to obtain a patent for a tool used only within our company?” However, even for internal tools, obtaining a patent offers tremendous benefits from a corporate strategy perspective.
① Establishing Barriers to Entry for Future “SaaS Sales”
A business efficiency tool that has delivered dramatic results within your own company is a system that competitors in the same industry would desperately want to get their hands on.An opportunity will inevitably arise in the future to package these tools as SaaS and sell them externally (B2B expansion). At that time, holding a patent will allow you to legally block imitation by large, well-funded companies and competitors, establish a dominant market position, and secure a new revenue stream.
② Preventing Patent Infringement Claims from Other Companies (Business Protection)
In fact, the primary purpose of obtaining a patent is to “safeguard your own business.” If you continue operations without securing a patent and another company later obtains a patent for a similar system, you run the risk of receiving a warning stating, “You are infringing our patent; either cease use or pay licensing fees” (proving prior use is extremely difficult).By obtaining a patent yourself, you eliminate the risk of being sued for infringement and can confidently drive forward your DX initiatives. This is referred to as a “defensive patent application.”
③ Enhancing Brand Power as a DX-Leading Company and Positive Impact on Fundraising
The fact that a company “holds its own patents for AI-powered business efficiency systems” serves as powerful, objective evidence to showcase its technological capabilities to the outside world.Not only does this significantly boost the company’s valuation when raising funds from investors and venture capital firms, but it also works to the company’s great advantage in recruiting top-tier IT engineers and data scientists by positioning it as a “leading company actively protecting cutting-edge technology.”
Chapter 5: Points to Note Regarding Generative AI Patents and Consulting with Patent Attorneys
Patent applications for systems utilizing generative AI involve specific considerations unique to the development process.
The Inventor Must Be a “Human”
Under current Japanese patent law, AI itself cannot be registered as the inventor. Even if AI writes the majority of the code, it is a human who defined the problem to be solved, provided appropriate instructions to the AI, and designed the overall system architecture. Therefore, the inventor listed on the application must be a “human (natural person),” such as a project manager or engineer who defined the requirements.Proper handling of rights in accordance with the company’s internal regulations on employee inventions is required.
Measures to Prevent Information Leaks (Loss of Novelty)
If you input your company’s confidential information or proprietary know-how into generative AI, there is a risk that it will be used as training data for the AI, inadvertently leading to information leakage (loss of novelty). It is important to strictly enforce opt-out (training refusal) settings or use a secure enterprise edition. Additionally, caution is required regarding violations of open-source software (OSS) licenses.
Selecting a Patent Attorney with Expertise in AI and IT
Patent applications for generative AI require a specialized approach that is entirely different from traditional patents for mechanical parts. When selecting a patent firm, use the following criteria: “Do they have a deep understanding of the latest IT and AI technologies (such as RAG, API integration, and vector databases)?” and “Are they familiar with the latest examination trends at the Patent Office?”
Quick Tip: Even without completed source code, a patent attorney can adequately assess the potential for patentability if you provide architectural diagrams or screen flowcharts that illustrate “how the system operates in terms of data flow.”
Summary
Here is a summary of the key points from this article.
- Even for systems that utilize generative AI, obtaining a patent is entirely possible if they incorporate unique innovations.
- The invention must satisfy the requirements of “novelty” and “non-obviousness,” and involve specific information processing using hardware.
- Patents are more likely to be granted for unique data processing (pre- and post-processing) or UI control, rather than the simple use of APIs.
- There are immense business benefits, such as monetization through SaaS, protection against competitors, and enhanced corporate value
“Could this AI dashboard we developed in-house be patented?” “I need to consult on securing patent rights quickly before competitors copy it.” If you’re an executive or DX manager with these concerns, please feel free to contact us.
Please also refer to our related articles on the deep connection between generative AI and patents, as well as the “multi-multi-claim” restriction.
Free Consultation on Patenting Generative AI Dashboards
It would be a great shame to give up, thinking, “Could this AI dashboard we developed in-house be patented?” or “It’s just an internal tool, so it’s probably not possible.” At the intellectual property firm EVORIX, patent attorneys with expertise in AI and software will review your system and provide free advice on the likelihood of obtaining a patent and the optimal strategy for securing your rights.We also welcome consultations even if your technical idea is still in the early stages of development.
#PatentAttorney #Patent #GenerativeAI #OperationalEfficiency #InternalDashboard #SoftwarePatent #AIPatent #DXPromotion #IPStrategy #SaaS
AUTHOR / Writer
Takefumi Sugiura
Representative Patent Attorney, EVORIX Intellectual Property Firm
Assists clients across a wide range of industries—including IT, manufacturing, startups, fashion, and healthcare—with everything from patent, trademark, design, and copyright applications to appeals and infringement litigation.He is also well-versed in intellectual property strategies for cutting-edge fields such as AI, IoT, Web3, and FinTech. He is a member of several organizations, including the Japan Patent Attorneys Association, the Asian Patent Attorneys Association (APAA), and the Japan Trademark Association (JTA).
