- Wide-ranging support for AI utilization, from high-precision AI chatbots to AI agent implementation -

Fujitsu today announced that it will offer its AI platform "Super RAG™," which enables highly accurate utilization of the large amount of proprietary knowledge accumulated within a company, in three new plans: "Basic," "Starter," and "Pro."

The currently available "Super RAG" service will be positioned as "Super RAG Basic," and starting today, "Super RAG Pro" will be available, which will include support for the design, development, and implementation of AI agents centered around RAG*. Additionally, the easier-to-implement "Super RAG Starter" service will be available from June of this year.

*Retrieval Augmented Generation technology that adds unique information, such as internal documents, to large-scale language models

https://cinnamon.ai/super-rag/

■ Background to the plan expansion

While the adoption of AI chatbots using RAG (Search Augmentation and Generation) has been expanding in recent years, many companies have only reached the proof-of-concept (PoC) stage, and the issue of not being able to establish them in actual business operations has become apparent. The main reason for this is that adoption is progressing without sufficient restructuring of business processes or clarification of the purpose of use, resulting in the mass production of "unused AI" that is disconnected from on-site needs.

Specifically, an increasing number of companies are facing more practical and structural barriers, such as "not being able to concretize the concept of business utilization," "the high level of expertise required to build and operate the RAG infrastructure," and "not being able to design and implement in-house AI agents that connect with business flows and core systems."

In light of this situation, we have redefined "Super RAG" as "the core of an AI platform that connects internal knowledge to decisions and actions." In addition to being used as a RAG platform for developing new AI agents, this platform also provides document processing functions that contribute to improving search accuracy through API integration without replacing existing RAG environments. It is designed to be flexibly used according to the purpose of use and development phase, from standalone use as an AI chatbot to implementing an AI agent by integrating with existing systems via API.

■ "Super RAG" ★3 new plans

We will develop the following three plans to match the phases in which companies utilize generative AI. Through optimal implementation support at each stage, we will continue to support the steady establishment of generative AI in business operations by linking in-house knowledge to "decisions" and "actions" in actual work.

①"Super RAG Pro" (available from today)
- Supporting the implementation of AI agents, the core of business automation -

"Super RAG Pro" is a service that uses "Super RAG Basic" as the RAG platform at its core and provides total support for the design, development, and implementation of "AI agents" that are deeply integrated with business flows and existing systems through API integration. In this service, AI agents go beyond simply responding to chats, making autonomous decisions and organizing information while referencing internal knowledge, and supporting specific business actions such as the following:

Main support contents

• ・Automatic import and utilization design of existing data (various files, RDBMS, etc.)
• ・Implementation of AI agents linked to the client's existing business systems, and support for the development of new UIs tailored to the client's needs
• ・Advanced search and decision-making processes using "Super RAG," which can be integrated with the existing RAG platform via API
• ・Support for streamlining and automation of business tasks (inquiries, applications, filing, information extraction, FAQ updates, decision-making, etc.)

●Implementation concept

• ・High maintainability: By not customizing the "Super RAG" itself and instead focusing on API utilization, you can enjoy product updates while maintaining stable operation and maintenance.
• ・Flexible scalability: For missing functions, we combine Dify and scratch development to implement dedicated agents optimized for individual tasks.
• ・Highly effective operation: We target projects with clear business objectives and expectations, and provide step-by-step support from PoC to integration into the production environment.

*This service is positioned to support the verification of implementation feasibility and implementation based on the business concept and data organized by the customer. It does not act as an agent for defining business requirements or organizing the business concept itself.

Example of system architecture for creating an AI for initial response to a disaster using "Super RAG Pro"

②"Super RAG Basic" (currently available)
- High-performance RAG platform at the core of company-wide AI agents -

"Super RAG Basic" is a core enterprise plan that can be used as a standalone, highly accurate RAG-type AI chatbot, as well as for API integration with external systems. It centrally manages corporate knowledge and acts as the core platform for AI agents, making it ideal for companies considering company-wide deployment or using large volumes of documents across multiple departments.

③"Super RAG Starter" (scheduled for release in June 2026)
- High-precision RAG now available for immediate use on a shared cloud -

"Super RAG Starter" is an entry-level plan that retains the core technology of the Basic plan while maintaining high security and reducing implementation costs by adopting a shared cloud configuration. It supports a smooth start by meeting the needs of those who want to "try it out on a departmental level first" or "want to conduct initial verification to plan how it will be used in business." After implementation, it is easy to gradually upgrade to Basic or Pro based on the confirmed effectiveness.

Main features

• High-precision index × High-quality generated AI answers remain the same
• You can quickly start using the shared cloud environment provided by our company.
• Cost-effective and optimized for small-scale departments and verification purposes

■ Pricing plan (planned)

Plan	Starter	Basic		Pro
Provided environment	Shared Cloud	Dedicated Cloud	Customer Cloud	Development support with Cinnamon AI
Data capacity	Approximately 10GB	Approximately 200GB		We will provide a quote each time depending on the work content and development scope.
Number of groups using the service	3 groups	Unlimited
Initial cost	700,000 yen	2.5 million yen	6 million yen
Basic fee	150,000 yen/month	300,000 yen/month
Query usage fee	10 yen per 2,000 tokens Minimum usage fee: 10,000 yen/month (2 million tokens/month)
Server costs	-	Approximately 120,000 yen/month	Customer prepared

*Prices and specifications for the Starter plan are tentative. Details will be announced in June 2026.
*In the case of a customer cloud environment, all external API costs, including LLM usage fees, will be borne by the customer.

■ "Super RAG" has an answer accuracy of over 90% without customization

Cinnamon AI's "Super RAG" is a proprietary AI product that allows you to build an LLM environment without tuning, analyzing complex documents such as tables and figures and generating highly accurate answers. Its greatest feature is that it goes beyond merely enhancing certain functions to "overall optimization" all processes, including preprocessing, search, and generation, for general purpose accuracy. In particular, during preprocessing, unlike conventional RAG, which mechanically separates documents based on the number of characters, Cinnamon AI's unique layout analysis technology automatically identifies titles and figures and creates optimal chunks while preserving the meaning and structure of the document. This allows it to highly structure unstructured data that is unsuitable for data utilization, enabling extremely accurate business automation using the RAG system.

"Super RAG" combines vector search and graph structure search to derive accurate answers without overlooking anything, even covering relationships not explicitly stated in documents. No prior data processing is required to optimize the search; simply register specialized information such as financial statements, technical documents, and internal regulations, and you can instantly create an AI assistant on-site with no-code operations in as little as five minutes. Its versatility allows it to be deployed to multiple use cases in a single environment, instantly providing "business-ready accuracy" in a wide range of areas, including responding to inquiries, creating reports, and small-lot, high-mix operations that require tacit knowledge.

In the latest update, we officially released a feature to integrate with Microsoft SharePoint. This allows companies without RAG expertise to safely and easily operate their own dedicated AI environment using existing internal storage. We also plan to integrate with Box in the future, further improving the efficiency of internal knowledge utilization and information sharing, thereby powerfully promoting corporate digital transformation. *Answer accuracy may vary depending on the type of document.

●Examples of using "Super RAG"

[Human Resources/Legal Department]Answering work regulations and checking contracts
[Sales Department]Product information confirmation, FAQ automation
[Manufacturing site]Knowledge sharing and safety management support
[Technology/IT department]Development knowledge search, business system support
[Management and Planning Department]IR information responses, management information search

*The assistant function also allows department staff to build and deploy the system.

Complex documents including figures and tables that can be used with "Super RAG" (image)

Comment from Cinnamon AI CEO, Hirano Miki:

"We believe that the value lies not in making AI smarter, but in enabling companies to act correctly, quickly, and without hesitation. Super RAG was born as a platform to convert knowledge lying dormant within companies into a form that can be used for decision-making and action. Now, in order to fully utilize that knowledge in actual business operations, we are stepping into the realm of AI agents, which drive the business itself. Rather than AI that follows trends, we will provide AI that continues to be used in the field and continues to change the will and actions of organizations."

■ Related pages

Cinnamon AI "Super RAG": https://cinnamon.ai/super-rag/　
Cinnamon AI "Super RAG" materials are available here Click here Please contact us at:

*"Super RAG™" mentioned in this press release is a trademark of Cinnamon Co., Ltd.