LegalTech Launches 'AI Pre-investigation Workflow' for Food Manufacturers on its AI Platform 'IPGenius'

LegalTech, Inc. has launched an 'AI Pre-investigation Workflow' for food manufacturers within its 'IPGenius' AI knowledge base for R&D and intellectual property.

This workflow leverages AI to organize patent specifications, experimental data, research notes, and table data from examples and comparative examples in the food sector, supporting the information organization essential for prior art and FTO (Freedom to Operate) searches.

Specifically, the workflow extracts the formulations and conditions evaluated in examples and comparative examples, and organizes ingredient names, product names, and abbreviations into generic names, chemical names, or broader concepts. This streamlines the creation of search term candidates, the organization of proprietary product structures, and cross-searches of past cases required before conducting investigations.

Background: The Burden of Pre-investigation Organization in the Food Sector

In R&D and intellectual property departments at food manufacturers, patent specifications and experimental data often contain diverse raw materials and formulation conditions, such as fats/oils, emulsifiers, additives, functional materials, powder materials, flavors, and stabilizers.

However, practical work involves tasks such as:

- Extracting actually evaluated formulations from examples and comparative examples.
- Organizing ingredient names, product names, abbreviations, and model numbers.
- Identifying generic names, chemical names, and broader concepts for search purposes.
- Decomposing the structure of proprietary products into a format comparable with third-party patents.
- Checking for past similar formulations or examination cases.

While these tasks are crucial prerequisites for prior art and FTO searches, they are often dependent on the experience and knowledge of individual staff, leading to inconsistencies in the granularity of organization and search perspectives.

Capabilities of the 'AI Pre-investigation Workflow' for Food Manufacturers

The newly launched workflow supports the following tasks:

1. Extraction of Formulations and Conditions to be Evaluated
It extracts evaluated formulations and conditions from examples, comparative examples, test examples, and table data in patent specifications and experimental data.

Examples of target information:
- Name of each ingredient
- Formulation ratio
- Units for formulation ratio
- Examples and comparative examples evaluated
- Evaluation items
- Evaluation results
- Differences in conditions

This enables organization focused on actually evaluated formulations rather than just candidate components listed broadly in the specifications.

2. Organization of Terminology (Ingredient Names, Product Names, Abbreviations)
In the food sector, a single component may be listed under multiple names, such as product names, abbreviations, common names, chemical names, or display names.

This workflow organizes naming candidates based on ingredient names in experimental data into categories such as:

- Generic names
- Chemical names
- Broader concepts
- Alternative names
- Similar ingredient names
- Substitute ingredient names

This prevents search omissions caused by searching only for product names or abbreviations and supports the creation of search term candidates for prior art and FTO searches.

3. Pre-processing for Prior Art and FTO Searches
In prior art and FTO searches, accurately decomposing the technology under investigation is crucial.

This workflow organizes food formulations, raw materials, applications, effects, and evaluation conditions, converting them into a format that is easy to use for investigations.

Application examples:
- Pre-investigation organization for new food materials.
- Formulation organization for fat/oil compositions, emulsified compositions, powder foods, and functional foods.
- Component organization for proprietary product FTO searches.
- Checking third-party patents when changing ingredients or formulations.
- Cross-searching past investigation notes and similar cases.

Difference from General Generative AI

While general generative AI tools like ChatGPT, Gemini, and Copilot can summarize or extract data from individual files, IPGenius differentiates itself by continuously accumulating R&D and IP data.

By combining this with operational prompts, RAG (Retrieval-Augmented Generation) search, source display, access control, and cross-searches of past cases, it serves as a knowledge base that can be reused within the organization. It is characterized by its ability to cross-search and organize past specifications, experimental data, investigation notes, and research notes, integrating them into the R&D and IP workflow, rather than simply reading files one by one.

About IPGenius

IPGenius is an AI knowledge base designed for R&D and IP departments.

It organizes accumulated internal R&D and IP data—such as patent documents, technical notes, invention proposals, experimental reports, quality data, and design data—into a format that is easy for AI to utilize.

Main application areas:
- Cross-searching past materials.
- Extracting and organizing invention candidates.
- Information organization prior to patent searches.
- Component organization prior to FTO searches.
- Patent list analysis.
- Converting R&D data into knowledge.

Future Prospects

LegalTech, Inc. will continue to sequentially expand AI workflows tailored to R&D and IP practices in various industries, including materials, chemicals, food, cosmetics, and electronic materials.