The engine operates by scanning the open web, product libraries, citation records, and competitor data simultaneously. Unlike standard generative AI tools, which often produce unverifiable citations or "hallucinated" products, the V3 system employs two distinct validation checks. The first confirms each candidate against a retrievable source, while the second filters results to ensure they align technically with specific patent claims rather than just broad subject matter.

During a recent soft launch at IPBC Global in San Diego, more than 20 licensing executives tested the platform using their own patent portfolios. Founder Bikram Singh noted that the primary concern among professionals was the potential for AI-generated inaccuracies. To address this, the system provides a transparency log, allowing users to review every product the engine considered, including those it rejected and the specific reasons for those exclusions.

The update introduces a Hit Matrix feature, enabling users to visualize entire portfolios in a single grid. This allows for immediate identification of high-value patents and potential targets, with the ability to export data directly to Excel for further analysis. V3 also supports batch processing, custom teardown report uploads, and filtering by company size, aiming to make deep patent research accessible to individual inventors and smaller firms alongside large-scale licensing operations.