ADF - Agentic Document Format

DraftprovisionalUS6/6 sections3 claims5 drawings

Configuration

Manage patent settings, bibliographic data, AI models, and invention disclosure

AI Model Configuration

Choose models for each task.

drafting Model

claims Model

analysis Model

Image Model

Patent Details

Core metadata and INID bibliographic fields

(54) Title

Patent Type

Jurisdiction

Entity Size

Docket Number

(21) Application Number

(11) Publication Number

(10) Kind Code

Technology Area

(51) CPC Codes

G06FH04LG06NG06Q

Inventors & Assignee

(72)/(75) Inventor details and (73) Assignee

Inventor 1

(73) Assignee

Correspondence Address

For PTO/SB/16 cover sheet

Invention Disclosure

Structured problem/solution and key features

Problem Being Solved

Solution / How It Works

The ADF (Agent Document Format) helps convert PDFs that are not easily readable by AI agents into a structured format that AI systems can understand and process effectively. Many PDFs contain complex layouts, embedded images, scanned text, or inconsistent formatting that makes it difficult for AI models to extract meaningful information directly. By transforming these documents into ADF, the content becomes clean, structured, and optimized for machine readability.

The ADF converts all document formats: text, docx, and other formats into an ADF standard format. It provides a downloadable player or reader so any person or any user can read these documents just like they currently do with a PDF. Any user with the ADF download is able then to manage, manipulate, change, and modify the ADF using an AI agent solution.

This conversion provides several key benefits. First, it significantly improves readability for AI agents, allowing them to interpret the document content more accurately. Second, it reduces token usage, because unnecessary formatting, duplicated metadata, and visual elements are removed or simplified. Lower token usage directly results in reduced processing costs when working with large language models.

Full Description

The present invention relates to a system and method for transforming non–machine-readable or poorly structured digital documents, particularly Portable Document Format (PDF) files, into an optimized Agent Document Format (ADF) designed for efficient processing by artificial intelligence agents. Many traditional PDFs contain complex layouts, embedded images, scanned text, and inconsistent formatting that hinder automated interpretation and extraction by AI systems. The invention introduces a conversion pipeline that parses document structures, extracts semantic content, and reorganizes the information into a standardized, lightweight, and structured representation that is directly consumable by AI agents and automated workflows. By normalizing textual content, removing redundant visual formatting, and structuring the document into hierarchical machine-readable components, the system significantly reduces computational overhead and token consumption during AI processing while improving accuracy and speed of document analysis. The resulting ADF enables AI agents to perform tasks such as information retrieval, summarization, reasoning, and data extraction more efficiently within agentic architectures where large language models interact with tools and external data sources to accomplish tasks. The invention therefore provides a scalable framework for converting legacy and complex document formats into AI-optimized data structures, enabling faster processing, lower operational cost, and improved interoperability across AI-driven applications, document management systems, and automated knowledge workflows.

Key Features

Known Prior Art

Related Applications

(60)/(62)/(63) Priority and related filings

Government Contract

35 U.S.C. 202(c)(6) disclosure

Made by a U.S. Government agencyMade under a U.S. Government contract

Filing Details

(45) Priority Date

(22) Filing Date

Status

Configuration

Manage patent settings, bibliographic data, AI models, and invention disclosure

AI Model Configuration

Choose models for each task.

drafting Model

claims Model

analysis Model

Image Model

Patent Details

Core metadata and INID bibliographic fields

(54) Title

Patent Type

Jurisdiction

Entity Size

Docket Number

(21) Application Number

(11) Publication Number

(10) Kind Code

Technology Area

(51) CPC Codes

G06FH04LG06NG06Q

Inventors & Assignee

(72)/(75) Inventor details and (73) Assignee

Inventor 1

(73) Assignee

Correspondence Address

For PTO/SB/16 cover sheet

Invention Disclosure

Structured problem/solution and key features

Problem Being Solved

Solution / How It Works

Full Description

Key Features

Known Prior Art

Related Applications

(60)/(62)/(63) Priority and related filings

Government Contract

35 U.S.C. 202(c)(6) disclosure

Made by a U.S. Government agencyMade under a U.S. Government contract

Filing Details

(45) Priority Date

(22) Filing Date

Status

ADF - Agentic Document Format

Command Palette

ADF - Agentic Document Format

Configuration

Configuration