Every patent starts as a problem someone couldn't ignore.

In 2025, SAS and JMP inventors earned 82 of them – spanning natural language processing, AI, fraud protection, IoT and more – each one rooted in the kind of friction that real organizations actually deal with.

At SAS' annual Patent Dinner, the focus was on the people who turned those problems into something protectable. Before the recipients received their patent plates, Vice President of Applied AI and Modeling Udo Sglavo joined the stage, drawing a connection between the evening’s celebrations and where SAS was 50 years ago. 

“A patent is never just one person doing smart work,” said Sglavo. "It's a team sport: co-inventors, reviewers, the patent team and the people who help turn an idea into something we can actually ship and support."

That framing matters. SAS has been solving enterprise problems since before it was formally a company – 206 people attended its first user conference in January 1976, drawn together by shared industry challenges. Five decades later, the same instinct drives its patent portfolio.

Before SAS was a company, there was already something else: a group of people with real industry problems, showing up, saying, ‘This matters. Keep going.' Udo Sglavo

Sglavo put it plainly: “SAS grows when we build what people genuinely need. SAS lasts when we build it with rigor. And SAS stays ahead when we turn real-world insight into outputs that only we can deliver,” he said. “That is what patents represent at their best.” 

That commitment to solving real-world problems still defines SAS inventors’ work today, Sglavo said. 

The 2025 patents reflect exactly that range. Innovations on Bayesian neural networks, causal graph analysis and hyperparameter search push the boundaries of what enterprise AI can do. Patents covering fraud detection, natural language processing and computer vision address problems businesses face right now. And work in synthetic data generation, bias mitigation in ML models and secure execution of analytical tasks signals where responsible AI development is heading.

Several patents also cross into less obvious areas, such as flood sensor allocation via distributed-parameter modeling, water quality monitoring and energy forecasting. It's a reminder that analytical rigor applies well beyond the data center.