Turn abandoned patents
into your own filing
into your own filing
This invention describes a software system designed to identify potential risks within contract clauses. It operates by learning from pre-labeled examples, called "seeds of text," where each example has already been assigned a specific risk category, such as "unlimited liability" or "missing protection." Once trained, the system can analyze new contract text to determine if its clauses fall into these predefined risk categories, and can also provide conversational answers to user questions about the contract.
Why it matters: Filed before the widespread adoption of powerful Large Language Models (LLMs), which can now significantly streamline the analysis of contract text, generate conversational responses, and classify clauses with greater accuracy and less custom engineering than was feasible in 2021. The core mechanism of learning from categorized text is now much more efficient to implement.
AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.
Reinvent this with AI
