• English
• 日本語

Business-Specific Generative AI Implementation Service

fusion AI / Two-Way RAG

fusion AI is a registered trademark of Synergy Research Corporation

Rapid Development and System Integration of Business-Specific Generative AI Capable of Communication with Internal and External Systems

• At first, you may be excited about the use of AI. But, are you getting tired of the continuous prompting and copy/paste?
• The reason for this is that the AI is not specialized for your business.
• When introducing business-specific AI, we face the following four problems and challenges. fusion AI / Two-Way RAG provides solutions to all of these.

Problems and challenges of business-specific AI,
and fusion AI / Two-Way RAG solutions

Problems and Challenges	Solutions
Learning cost	· Low-cost implementation using Open AI’s latest API technology · No need for LLM (Large Language Models) development or fine tuning
Business-specific user interface	· A wide range of user requests can be expressed in language · Customizable chatbots for realizing proprietary UIs.
Connection to internal and external systems	· Multiple options can be selected depending on the environment (see diagram below). · With method (3), you can enter in the shortest possible lead time.
Introduction Process	· Providing processes from requirements analysis to proto development & SI · Global project support

How fusion AI / Two-way RAG works

explanation

• We introduce assistants for each business area where AI utilization is desired.
• Assistants serve as an easy means for constructing business-specific generative AI.
• To realize the assistants, we use the Open AI Assistants API, the latest API (Application Programming Interface) from Open AI, a leader in synthetic AI.
• For the platform of choice for the AI generation, you can choose between Open AI’s platform and
• Microsoft Azure Open AI (AOAI).
• fusion AI / Two-Way RAG offers a specialized chatbot. The chatbot can be configured with multiple assistants to choose from, and custom-developed with tabular, image, and graphical screen elements to meet the user’s requirements.

• fusion AI / Two-Way RAG provides the two-way data exchange with internal and external systems required for business-specific AI in three ways
• In method (1), the target system is connected via API.
• In Method (2), a database management system (DBMS) is introduced to realize data transfer with AI, and then the connection to the target internal and external systems is carried out. Alternatively, a new system for AI-related operations will be built around this database.
• In Method (3), the task management system REDMINE is introduced and connected via API. With this method, it’s possible to input data from the task management system into the generative AI, or vice versa, allowing for system construction with minimal investment and time. Additionally, by using the task management system, it’s possible to achieve business collaboration between generative AI usage tasks and general tasks.

• RAG (Retrieval-Augmented Generation), which is considered one of the means to achieve business-specific generative AI, cannot realize the bidirectional data flow mentioned above. Furthermore, in typical RAG systems, real-time searching of business data is not possible.
• It is not possible to achieve data flow in one direction. Also, RAGs usually do not allow for real-time retrieval of business data.
• fusion AI / Two-Way RAG essentially assumes real-time data flow to satisfy the requirements of general generative AI usage.

• While generative AI can be used in major languages and translations between them, REDMINE, as mentioned above, is already multilingual (supporting over 50 languages, including Japanese). Therefore, according to Method (3), it’s possible to swiftly launch international generative AI usage tasks.
• Since REDMINE is open-source, for simple requirements, even deployments with tens of thousands of users can utilize the software license-free.

Q&A page is here

Application Areas

It can be applied to all areas of work where texts, data and profiles are generated, transformed, edited and summarized using a generative AI. We are currently preparing a collection of use case hints and use case templates for a wide range of applications (Press release coming soon).

Usecase：　Customer complaint handling

Introduction Process

fusion AI / Two-way RAG responds quickly to the unique requirements of each organization. For this reason, we offer free individual consultations, consulting and requirement analysis phases.

The lead time from consulting and requirements analysis to development is assumed to be one month per business in the case of a standard use case using the above-mentioned method (3). In this case, the price is 450,000 yen. In the case of method (1) and (2), API research and system development fees are required separately.
For high-level requirements, we will sign a contract for consulting and requirements analysis only, and after the work is completed, we will present detailed development plans and opinions.
Usage Support and Maintenance Contracts are quoted based on system environment and service level. Please contact us for more details and pricing.

Use Case Templates

Use case templates are being developed for typical AI applications.

Q&A

Q&A page is here

For inquiries, please contact

For further information, please contact us.

Synergy Research Corporation

fusion AI is a registered trademark of Synergy Research Corporation.

2024/05/17