Computing US 12,572,575 B2 Not in force

Using large language models to generate search query answers

This invention describes a system that answers user questions using a large language model (LLM). It works by first breaking down stored content into smaller text sections. When a user asks a question, the system finds the most relevant small text sections and sends them to an LLM, which then generates an answer by summarizing those sections.

Why it matters: Filed in 2024, when LLMs were rapidly advancing, the specific techniques claimed (like text chunking and summary generation) have become standard features in LLM frameworks by 2026, making implementation easier but potentially reducing distinctiveness.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedSoftware logic, LLM integration, content processing

3 AI reinvention directions for this patent

See all three

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

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

+10