AutoPat outperforms human patent searchers in two Unified Patents contests.
Autonomous prior-art searches produced winning submissions in two consecutive invalidity contests—surfacing references missed by professional participants.
02
Consecutive winning contests
100%
Autonomous search process
Dozens
Of competing submissions
In September 2025, prior art identified through AutoPat contributed to winning submissions in two consecutive crowdsourced invalidity contests run by Unified Patents.
Unified's PATROLL contests invite researchers and search firms to find previously uncited prior art against patents asserted in high-stakes litigation. Unified reported dozens of submissions across the challenges. PatentPlusAI executed its searches using AutoPat, which navigated and evaluated the global prior-art corpus autonomously rather than relying on human keyword curation.
Two difficult prior-art challenges
Contest 01
U.S. Patent 9,450,956
$2,000 award shared by three winning participants
PatentPlusAI was named alongside Tanya Chauhan and Vivek Yadav for prior-art submissions addressing a Proxense authentication patent.
Contest 02
U.S. Patent 11,095,640
$3,000 award for the winning submission
PatentPlusAI was awarded the winning submission for prior art addressing claim 17 of a second Proxense authentication patent in the Cloud Native Heroes Challenge.
Why the result matters
Patent searching rewards persistence and semantic breadth. A thorough searcher must interpret claim language, translate between diverse technical vocabularies, traverse citation graphs and patent families, and formulate new search hypotheses after obvious query strategies are exhausted. AutoPat is architected to sustain that depth across global patent and scientific literature.
The winning references were identified entirely through AutoPat's autonomous research workflow. These results provide concrete empirical validation in open competition: the system surfaced prior art strong enough to win alongside and ahead of submissions from experienced professional searchers.