AI-Supported IP Assessment: University of Minnesota

The Arizona State University Artificial Intelligence Cloud Innovation Center, powered by Amazon Web Services (AWS), collaborated with the University of Minnesota to explore the use of artificial intelligence in supporting early-stage evaluation of invention disclosures. Academic research institutions review thousands of intellectual property submissions each year, requiring extensive literature searches, novelty assessments, and commercialization analysis. The prototype solution aims to streamline the assessment workflow by assisting reviewers with AI-supported insights while maintaining human oversight and institutional decision-making standards.

The University of Minnesota’s current intellectual property evaluation process is labor-intensive and requires coordination across multiple teams. Reviewing invention disclosures often involves manual literature analysis, expert judgment, and iterative committee review — which can slow timelines, limit throughput, and introduce variability in evaluation criteria. The university sought a solution that could help teams accelerate the assessment process while preserving fairness, accuracy, and academic rigor.

The ASU AI CIC team designed the proof-of-concept solution to align with the university’s existing intellectual property evaluation workflow. The prototype focuses on supporting reviewers during the early stages of assessment by using AI to identify similarities in patent literature, extract and organize key details from invention disclosure forms, and generate preliminary commercial assessment summaries. A web-based interface allows users to review results and track evaluation progress.

While the tool assists with research and early analysis, it does not draft patent applications, make legal or patentability determinations, or integrate directly into the university’s production systems. Human judgment and institutional review processes remain central to decision-making.

The prototype demonstrates how artificial intelligence can support large-scale research commercialization efforts by reducing manual review workloads, improving consistency in evaluation criteria, and allowing expert reviewers to focus on high-impact decision-making instead of repetitive screening tasks.

"Our pilot with the Artificial Intelligence Cloud Innovation Center (AI CIC) and the Patent TPMs proved that we can significantly cut the time and effort it takes to move from invention disclosure to an initial recommendation. While the AI-generated reports aren't yet at our full quality standard, the efficiency gains are undeniable. We’re encouraged by the trajectory of these models and are moving toward a process where GenAI handles the heavy lifting of the initial assessment, followed by a human expert review. This shift is about providing better customer service to our inventors and taking a more disciplined approach to our time and labor expenses."

- Kevin Anderson, Licensing Director, University of Minnesota 

This approach is not limited to a single institution. The framework can be adapted for use in university technology transfer offices, government research labs, corporate R&D teams, innovation incubators and licensing organizations. Because it can be aligned to different evaluation rubrics, document formats and review workflows, the system offers a flexible model for organizations that manage complex intellectual property assessments.