Bellsystem24 Completes Demonstration Experiment Using Google Cloud Generated AI with Microsoft Japan to Commercialize

Background of Initiatives

It has been estimated*1 that Japan will have a labor shortage of approximately 3.5 million workers by 2030, and companies are expected to face the challenge of securing more workers in the future, which will lead to an increase in outsourcing needs by allocating human resources to core operations and reducing working hours through reforms in work styles.
In addition, with the recent evolution of technology, such as the emergence of generative AI, it is expected that the hybridization and automation of contact centers will rapidly advance. Against this backdrop, Bellsystem24 conducted a demonstration test of generative AI in contact center operations.
*1Source: Recruit Works Research Institute, "Works 179 Labor Shortage Society vol. 3: Reskilling Advanced Countries: Their Vision and Current Location Labor Market Future Estimates 2030".

Bellsystem24 has so far established the Innovation & Communication Science Laboratories (ICSL) jointly with Sony Computer Science Laboratories, Inc. with the aim of upgrading contact centers through the use of AI.

ICSL has developed its own AI search engine "Mopas®" to support reduction of e-mail operations and "Knowledge Creator®" to enable on-site knowledge maintenance of FAQs, etc. ICSL has been engaged in joint research and development of an AI engine to build a next-generation contact center by combining its long accumulated operational know-how in the field of contact centers and new technological domains such as AI, natural language processing, and We have been conducting joint research and development of AI engines for the construction of next-generation contact centers that combine our long-accumulated operational know-how at contact centers with new technological domains such as AI, natural language processing, analytics, and big data. In addition, we have promoted the use of AI in contact centers, such as by already utilizing the contact center automation AI chatbot and AI voicebot "ekubot®" in our operations.

We have established a cross-functional project in the area of generative AI, and are conducting various studies to develop new services that ensure efficiency and security, aiming at contact center automation using generative AI in order to create a new contact center business.

Process for demonstration tests and future commercialization

In this verification experiment, the strengths of generative AI, such as dialogue data summarization, FAQ creation, and VOC excavation, were used to verify the system based on dialogue data from actual, anonymized calls. As a result, we were able to demonstrate the effectiveness in reducing the burden on operators by summarizing dialogue data with high accuracy and significantly reducing the processing time required per call.

For future commercialization, based on this knowledge and know-how, we will summarize dialogue data ranging from general inquiries to industry/industry-specific inquiries, and accumulate base knowledge while tuning the AI.

Based on this knowledge, we will promote the automation of contact centers by developing AI specialized for each industry. Furthermore, we aim to develop "AI application development," which will be responsible for customized AI response for each client company, and "AI knowledge base," which will allow AI to continuously learn. With the support of Microsoft Japan and Google Cloud, we plan to proceed through certain processes such as requirement gathering, business design, selection of appropriate technologies, and development and operation of prototypes.

In the course of such commercialization, we will continue to develop human resources with both operational business operation skills and IT skills to handle data appropriately, accumulate know-how and expertise in "data management" for data collection and utilization, and also aim for business beyond that.

Bellsystem24 will continue to strive to solve social issues through services that utilize AI and other new technologies as part of the social infrastructure, with proper understanding of technology, safety and ethics, and consideration for personal information.

About Azure OpenAI Service

Azure OpenAI Service, developed by Microsoft Japan, is one of the generated AI services that Microsoft offers to enterprises, using technology from OpenAI, a company that develops AI.
Large-scale language models such as GPT-3.5 and GPT-4 models are available on Microsoft Azure, Microsoft's cloud platform, and comply with Microsoft Azure's security and compliance standards, providing security features such as data encryption, access control, and audit logs. security features such as data encryption, access control, and audit logs.
https://azure.microsoft.com/ja-jp/products/ai-services/openai-service/

About "PaLM2" of "Vertex AI

PaLM2, available on Google Cloud's Vertex AI machine learning platform, is a state-of-the-art large-scale language model with improved multilingual, inference, and coding capabilities. PaLM2 performs well in natural language processing tasks, with the ability to understand the meaning and context of a language. It can also efficiently analyze large amounts of textual data through its fast inference processing, providing superior performance in parsing and understanding the meaning of textual data.
https://japan.googleblog.com/2023/05/palm-2.html