Global Site
Breadcrumb navigation
Interview
Advancing NEC’s Unique Approach to IP DX with Generative AI
September 29, 2025
Generative AI is widely used across many sectors of society. However, concerns regarding the reliability of AI-generated answers, security risks, and the significant human and financial resources required to develop AI tailored for specific business operations still remain. In response to these challenges, NEC’s Intellectual Property Division launched a project in 2024 to introduce generative AI solutions into its operations. By working closely with experts both inside and outside the company, the project team succeeded in deploying these solutions within the company. We spoke in depth with one of the core members of this team to learn how they have transformed IP operations through digital innovation at NEC and what their vision is for the future.
Profile
Drawn to NEC’s commitment to technology and its wide-ranging business activities—from submarine cables and space to agriculture and drug discovery—Kenichi Ueda joined NEC in February 2023. Mr. Ueda works on supporting new business development from an intellectual property perspective and, with support from NEC’s CEO, CTO, and the Head of the Intellectual Property Management and Rule Making/Standardization Division, has been actively leading the digital transformation of IP operations (also known as “IP DX”). He has also been involved in efforts to commercialize technologies developed in-house, with the aim of create meaningful value for society through his ongoing work.
Kenichi Ueda
Senior Intellectual Property Analyst
Intellectual Property Planning and Strategy Group
Reducing the Time Spent on Specialized Intellectual Property Tasks
Could you tell us about NEC’s approach to digital transformation in intellectual property?
We are harnessing generative AI to make our intellectual property operations more efficient. Since much of IP work involves handling and analyzing large volumes of text, it is the perfect usage case for large language models (LLMs), which are a leading technology in generative AI. In fact, we have already implemented and are using three solutions to reduce work time internally.
The first is a solution that supports inventors with generating ideas and drafting patent applications. Typically, when inventors prepare to file a patent, they start with a simple idea sheet. By inputting this sheet into an LLM and using the model to expand and refine the concepts, inventors can identify broader applications. For example, if an inventor initially only considered using their technology in airport network cameras, the AI might suggest additional applications in places such hospitals or schools by making minor adjustments. These insights can help inventors expand the scope of rights and lead to stronger patent applications.
The second is a solution that enhances prior art searches. As an invention cannot be patented if similar prior art exists, inventors must conduct research into whether similar intellectual property exists prior to filing their own applications. However, this step has traditionally been a complicated and technical process. As the research is highly specialized and advanced, many inventors have found it difficult to adequately review the relevant literature themselves. To solve this, we utilize Retrieval-Augmented Generation (RAG) technology, which converts past patent documents into vectors. Inventors can then simply describe their inventions in natural language, while AI automatically searches for documents with similar vectors, making prior art searches more efficient and accessible. We have also developed an AI that is trained on past decisions made by patent office examiners, offering inventors immediate feedback on whether their ideas are likely to meet patent requirements.
The third is a solution that supports a step known as "patent prosecution," which involves receiving and responding to feedback from the patent examiners after an application has been submitted. This AI technology helps inventors assess the validity of examiners’ comments and assists them in preparing effective responses and counterarguments.
All of these solutions are already in use at our company, and have significantly reduced the time required for these specialized tasks.
Identifying New Uses for Unused Patents and Transferring Them to Startups
It appears that AI has had a major impact on making your IP operations more efficient.
Yes. However, AI is not solely about increasing efficiency. We have continuously explored ways for our IP division to contribute new value that enhances corporate growth and help solve societal issues. One outcome of this effort has been our initiative to use AI to identify new applications for patents that were not being used internally. To date, three of these unused patents have been transferred to startups, enabling their technologies to be put to practical use in society.
Our system uses generative AI to search for potential new applications for unused patents. At the same time, in collaboration with financial institutions, we build databases of startups supported by those institutions. The system then matches potential application fields with startups, enabling efficient identification of potential uses for underutilized patents. We simply enter the patent number of an unused patent into the system, and the AI generates ideas for possible business applications and suggests relevant startup partners.
For example, among the cases mentioned earlier, the AI suggested that an image recognition technology that had not been used internally and previously only considered for use in manufacturing could instead be used for rehabilitation assessment in the healthcare field. We were surprised to see new possibilities emerge from such a different perspective.
Based on this suggestion, we partnered with Parrots Inc., a startup developing systems to monitor the behavior of people with Alzheimer’s disease, and transferred the patent to them. In this way, we were able to uncover new value from intellectual property that had previously been unused.
Building a Shared SaaS Platform to Support the IP Divisions of Other Companies
Are these IP DX initiatives only being carried out within your company?
Not at all. We are currently expanding these activities in collaboration with the IP divisions of other companies as well (over 20 as of September 2025). At NEC, we have a concept called "Client Zero," which refers to the practice of introducing the latest cutting-edge technologies internally first and treating ourselves as the initial client in order to refine and validate our solutions. While we initially focused on promoting IP DX within NEC, conversations with other companies revealed that many of them face similar challenges. Using the expertise and improvements we have accumulated, we are now working to offer our solutions to other organizations’ IP departments as well. We are currently running multiple proof-of-concept (PoC) projects with partner companies and plan to continue making enhancements based on their feedback.
One common theme I hear from discussions with various companies is that, although they want to pursue IP digital transformation, the required investment in resources and equipment is a significant hurdle. As they have expressed, advancing IP DX can require substantial costs, both in terms of personnel and equipment. As such, as a developer of AI and provider of IT solutions, NEC aims to build a shared foundation, a shared SaaS platform, to help address these challenges and make it easier for IP divisions across Japan to advance their digital transformation efficiently.
Fortunately, we have robust technical capabilities and deep industry expertise. In terms of technology, we have extensive experience utilizing generative AI and have even developed our own AI models in-house. We hold weekly meetings with researchers working on "cotomi," NEC’s proprietary large language model, to keep technological development moving forward. We also collaborate extensively with both internal and external vendors to deliver cutting-edge solutions.
Additionally, security is another of our major strengths. Because we handle highly confidential intellectual property information, NEC has built an operational system that complies with high security standards, and our key advantage lies in our ability to maintain rigorous management systems that ensure a safe and secure environment from development to operation.
Another major strength is our potential for future growth. For example, NEC possesses world-class image recognition technologies and, in the IP field, analyzing diagrams and drawings is just as important as analyzing text. We are actively researching how to apply AI to diagram analysis and are confident that there is much more that can be achieved through further AI innovation.
Achieving NEC’s Purpose Through Intellectual Property
What are your goals for the future?
I believe that intellectual property digital transformation plays an essential role in realizing NEC’s purpose: “to create the social values of safety, security, fairness, and efficiency, and to build a more sustainable world in which everyone can reach their full potential.” By utilizing generative AI in a secure environment and enabling natural-language interaction with AI to support IP operations, we aim to ensure that these tasks fairer and more efficient for everyone, while also enabling employees to focus on more creative and human-centered work.
As such, I plan to remain actively engaged in these efforts to continue creating social value. I originally became interested in joining NEC because I wanted to support new business development across a wide range of fields through intellectual property. From that perspective, treating IP DX as a new business initiative and providing unique value to the IP departments of other organizations is not something typically experienced in a conventional IP role, and I find the work especially rewarding and meaningful.
With strong support from senior management—including from our CEO and CTO—NEC is able to drive IP DX as one unified organization, which is uncommon. I hope to make full use of this environment and continue contributing to work that creates lasting value for society.
- *The IP DX solutions discussed in this article are designed to support operational activities and are not intended to constitute legal advice.
- *Company names and job titles are as of the date of the interview.
