Konica Minolta IP Division runs ~500 generative AI prior art searches monthly

Konica Minolta’s Intellectual Property Division needed to improve productivity across IP-related work for R&D departments, the IP Division itself, and domestic patent firms. High-frequency tasks such as prior art patent searches and technology trend surveys required repeated query formulation and manual review of prior documents, slowing invention refinement and filing decisions.

The IP Division developed and deployed a suite of in-house AI applications built on Microsoft Azure OpenAI, accessed via an intranet “IP Apps Portal.” A flagship app, “Smart AI Prior Art Search,” lets users enter an invention idea in natural language; the system extracts similar patent documents and relevant paragraphs, then uses generative AI to compare, evaluate, and advise on improvements and filing considerations. The architecture runs as a four-block web application (prior art search, screening, invention evaluation, and AI chat) on an internal on-premises server, with companion tools such as “IP Landscape on Demand” for information analysis.

Since company-wide release in March 2025, Smart AI Prior Art Search has been executed approximately 500 times per month and has reduced the workload required for prior art searches. The applications are used daily for prior art searches, technology trend analysis, and patent specification drafting support, improving productivity. The IP Division continues to refine the tools from field feedback and plans to expand AI use into clearance searches, continuous monitoring, and specification quality improvement.

• • Start with daily AI usage and a “try it out” culture, then move stepwise toward measurable value creation.
• • Prioritize high-frequency IP tasks (prior art search and tech trend analysis) when rolling generative AI out company-wide.
• • Combine document retrieval/screening with generative evaluation and chat so users refine invention ideas while understanding prior art.