A summary of the latest examples of companies using generative AI! A thorough explanation of the benefits and implementation methods.

In recent years, generative AI tools such as ChatGPT and Gemini have appeared one after another and are being implemented across all industries. However, when it comes to actually introducing them to your own company, some may have questions such as "Can we really get results from the implementation?" or "How can it be utilized?"

Therefore, in this article, we explain cases where companies have actually implemented generative AI. The content provides a concrete view of the effects obtained in actual business scenes, implementation methods, and how to consider the costs involved in implementation.

If you are seriously considering the utilization of generative AI in business, please read to the end.

1. Major Business Uses of Generative AI

Generative AI is a type of AI technology that is mainly based on deep learning and automatically generates text, images, code, etc., based on vast amounts of data.

Specifically, it can be divided into the following categories based on what can be generated.

In recent years, multimodal generative AI, such as Google's Gemini which can handle both text and images, has also appeared.

In addition, gene

rative AI has a wide range of uses and can be utilized in the following fields.

With a single generative AI, you can achieve task automation in a wide range of fields, from marketing to customer support and creative work.

2. 11 Latest Business Implementation Cases of Generative AI

Here, we introduce the latest business implementation cases of generative AI.

・Achieving operational efficiency with task-specific generative AI (Circulation)

Circulation Co., Ltd., which supports solving corporate issues by utilizing the experience and insights of professional personnel, was spending a lot of time in its sales department on organizing information, processing data, and searching for information to understand customers.

Additionally, in the back office, inefficient business processes were a major challenge, such as being required to respond to hundreds of internal inquiries every month.

Therefore, to improve business efficiency, they made an inquiry to "Alli LLM App Market," an all-in-one generative AI and LLM application platform provided by Allganize Japan Co., Ltd., and began internal verification experiments.

As a result, in the sales department, the creation and customization of generative AI apps progressed spontaneously, and business efficiency was achieved. It also serves as an auxiliary tool for education and training, such as checking talk scripts.

Regarding implementation costs, flexible pricing that is easy to scale—with an initial cost of 300,000 yen and a running cost starting from 300,000 yen per month under a license system with an unlimited number of users—seems to have encouraged the implementation.

Reference: ■Implementation Case■ [Circulation Co., Ltd.] Realizing a dramatic improvement in productivity by releasing dozens of generative AI apps while ensuring data privacy

・Supporting internal IT help desk with generative AI chatbot (OPTAGE)

OPTAGE Inc., a member of the Kansai Electric Power Group, introduced "OPTAGE Generative AI Chat," a generative AI environment for all approximately 2,900 employees, in July 2023.

Then, as the need for knowledge search within the company increased, the following points became clear through a PoC utilizing a generative AI chatbot for the internal IT help desk.

To address these technical challenges, the company requested technical support for utilizing generative AI from Classmethod, Inc., aiming for more advanced operations. As a result of the verification, they achieved solid results, such as an answer accuracy exceeding 80% when referencing internal documents in a specific verification method.

Reference: Supporting internal inquiries by introducing a generative AI chatbot; also contributing to business improvement efforts and accumulation of AI utilization knowledge

・Streamlining inquiry response from stores to headquarters using generative AI chatbots (Kyoto Toyopet)

At Kyoto Toyopet, which handles the sales and maintenance of Toyota and Lexus vehicles, inquiries to headquarters occurred frequently when engineers performed repair work based on manufacturer warranties or national recall systems.

However, responses from headquarters relied on the experience and knowledge of specific personnel, and the handover of personalized information was a challenge.

Therefore, to reduce the man-hours required for responding to inquiries from employees at each store to headquarters and for searching for materials, they decided to introduce "SELFBOT," a generative AI chatbot provided by SELF Inc.

SELFBOT is capable of high-precision contextual understanding and responding to questions in free-input format. Additionally, bots can be created for each department. It is evaluated for its ability to be implemented at a low price and its capacity to handle future functional expansions.

After the implementation of SELFBOT, the display of documents registered for learning in the bot enabled employees to quickly access the materials and information they were looking for. In addition to significantly streamlining inquiry responses from stores to headquarters, it also led to solving challenges related to the handover of personalized knowledge.

Reference: Streamlining store to headquarters inquiry response with "Generative AI Chatbots"! Kyoto Toyopet's Efforts [Implementation Case Interview]

・Reducing information collection work by approximately 40% (Japan Airport Consultants)

Japan Airport Consultants, Inc., Japan's only airport specialist consulting firm, was manually collecting, translating, and organizing information about domestic and international airports and aviation, which could sometimes take several months. Therefore, it was difficult to devote sufficient time to the core business of planning and development, necessitating countermeasures.

While exploring ways to improve business efficiency, they focused on and decided to adopt "Robo-Resa" provided by Mitsubishi Research Institute, Inc. With the introduction of Robo-Resa, they utilized the automatic website information collection function to automatically deliver update information for the latest articles.

Additionally, with the article summarization function utilizing generative AI, they became able to instantaneously grasp necessary information without spending time and effort.

As a result, they succeeded in reducing man-hours by approximately 40% for information collection work as a whole. In particular, they succeeded in streamlining the tasks of grasping the latest trends in the overseas aviation industry and organizing/analyzing information, which had taken an enormous amount of time, allowing them to concentrate on their core planning business.

Reference: Implementation case of generative AI solution "Robo-Resa"

・Rapidly producing event reports and SNS posts from event audio data (Enageed)

Enageed Co., Ltd., which develops inquiry-based learning programs for junior and senior high school students and training services for companies, faced challenges in both human resources and budget for PR and marketing content production. A lack of resources for personnel meant they were continuously preoccupied with daily production tasks, making it difficult to formulate essential PR strategies.

To solve these challenges, they adopted "Bakusoku AI Writing," a generative AI content production service provided by unname inc.

With the introduction of Bakusoku AI Writing, they became able to rapidly generate first drafts of event reports based on event audio data. Additionally, they introduced a mechanism to automatically create multiple patterns of SNS posts, such as proposal types and service introductions, by utilizing existing materials.

As a result, they succeeded in significantly reducing the man-hours for content production tasks. Furthermore, through the utilization of generative AI, they were able to acquire new marketing and PR methods and ideas, enabling more multifaceted information dissemination than before.

Because a free trial is available for "Bakusoku AI Writing," it is possible to make realistic judgments when internally considering implementation costs and implementation effects.

Reference: Enageed Co., Ltd. delivers high-quality first drafts of event reports at "explosive speed" from audio data materials.

・Providing high-impact advertising video production services (Hakuhodo DY Holdings)

Hakuhodo DY Holdings Inc. and its group company, i-rep Co., Ltd., provide "H-AI EYE TRACKER," an advertising video production service utilizing independently developed eye-tracking AI technology.

Advertising videos can convey more information than traditional static images or text ads. On the other hand, methods for measuring and analyzing results have become more complex.

In particular, there are challenges in production processes that effectively capture both customer acquisition and awareness/favorability formation, and more precise creative improvement is urgent.

In H-AI EYE TRACKER, an AI that has learned human gaze tendencies from eye-tracking data predicts and visualizes the areas where gaze is concentrated within a video in heat map format. Then, it creates improved content based on those results.

As a result, user surveys and AB testing that were previously necessary are no longer required, and a more efficient creative improvement process can be constructed.

Reference: Start of providing advertising video production service "H-AI EYE TRACKER" utilizing Hakuhodo DY Group's unique AI technology

・Automating routine tasks using Excel and Word (Honda)

Honda, a major transportation equipment manufacturer, introduced "Microsoft Copilot" company-wide from October 2023 to promote the automation of routine tasks using Excel and Word. Microsoft Copilot is a generative AI tool provided by Microsoft that supports the creation of emails, Word documents, etc., using AI.

After the introduction of Microsoft Copilot, tasks for automatic technical document creation that previously required 3 hours were shortened to 15 minutes, achieving a time reduction of 17 hours per person per month.

In the design department, they succeeded in reducing design change rework by 42% by having the AI automatically check the consistency between 3D models and specification documents.

If the Microsoft Office environment is already utilized within the company, adding Copilot, a generative AI function, eliminates the need to build new infrastructure or systems. Additionally, because it can be introduced by adding to existing licenses, implementation costs can be kept low.

Reference: Honda accelerates the second founding period; the reasons for the decision and the large-scale, speedy introduction of Copilot for Microsoft 365 that changes work

・Automatic generation of product descriptions (Mercari)

On October 17, 2023, Mercari, Inc. began providing "Mercari AI Assist," an AI assistant service for sellers. Mercari AI Assist is a tool that proposes solutions for questions about listing or purchasing and encourages optimal actions when using Mercari.

With the "automatic product description creation function" implemented as the first-phase function, the average time to complete a listing can be shortened from 15 minutes to 2 minutes.

Additionally, products that adopted "AI Propose," which proposes improvements to product information based on Mercari's past information, had a 43% higher closing rate than non-adopted products, and specifically in the used home appliance category, they succeeded in shortening the average sales period by 3.7 days.

Reference: Mercari begins providing "Mercari AI Assist," which utilizes generative AI and LLM to encourage optimal actions for customers

・Chatbots automating inquiry responses (Bank of Yokohama)

The Bank of Yokohama released "AI Help Desk for Microsoft Teams," which utilizes deep learning and NLP (Natural Language Processing), to all bank employees from early 2024. Developed by PKSHA Workplace Inc., "AI Help Desk" is a service that realizes the automation of internal help desks on Microsoft Teams by coordinating the following functions:

The background for the implementation is related to the prediction of a surge in internal inquiries following the introduction of the next-generation SFA/CRM and loan screening system. Inquiries that could not be handled by an AI chatbot alone were anticipated, and a mechanism was needed to seamlessly coordinate from AI to human response.

Therefore, when the AI Help Desk was introduced to primary inquiries in inquiry response tasks, it reduced FAQ search time by 85%, and specifically, inquiry response efficiency for legal matters improved threefold. It also demonstrated effects outside of inquiry response tasks; in loan screening, it succeeded in commercializing 38% of new products proposed by the AI.

Reference: Bank of Yokohama introduces "AI Help Desk for Microsoft Teams"

・Supporting basic building design (Obayashi Corporation)

Obayashi Corporation, one of Japan's leading general construction companies, jointly developed "AiCorb" with SRI International in Silicon Valley, USA, to streamline architectural design.

In traditional architectural design, designers manually created design proposals using sketches and CAD and had to reconsider them many times to match customer requests.

However, AiCorb has learned a wide range of design patterns and can instantaneously generate facade designs from sketch data relative to building outlines. Furthermore, by coordinating with "Hypar," a platform for designers, an integrated design process including building volume design can be realized.

As a result, rapid consensus building with customers was realized, and basic building designs can now be created in one-tenth the time. Additionally, more than 200 variations can be generated from a single basic concept, achieving a customer satisfaction rate of 92%.

Reference: Developed "AiCorb®" to streamline work in the early stages of architectural design

・Drug discovery based on unique protein structure analysis AI (Fujitsu)

Fujitsu Limited and RIKEN began joint research in May 2022 and, in January 2023, developed a new protein structure analysis AI using generative AI.

Based on a large number of electron microscope images of target proteins, the protein structure analysis AI generates the three-dimensional structures the protein can take over time and their proportions. Then, it predicts structural changes in the target protein from the estimated proportions.

Ultimately, structural changes in proteins can be reproduced as continuous deformations of 3D density maps. With this technology, the discovery speed of drug discovery candidate substances by protein structure analysis AI has improved more than tenfold compared to conventional methods, and it is expected to contribute significantly to speeding up the design process for drugs that bind to targets such as bacteria and viruses.

In 2024, two of the compounds proposed by the AI reached the clinical trial stage, and the fact that new drug development is being realized at one-fifth the traditional development cost is attracting attention as a major result.

Reference: Fujitsu and RIKEN develop drug discovery technology based on unique generative AI; realization of predicting wide-range structural changes in proteins from electron microscope images

3. How to Implement Generative AI in Business

The specific steps for implementing generative AI in business are as follows:

By proceeding with careful verification at each step, you can increase the implementation effect of generative AI.

4. Summary

As seen in the implementation cases introduced this time, generative AI is utilized in a wide range of business areas such as support tasks, advertising production, and information collection in companies, supporting operational efficiency and creative output.

However, to ensure the implementation effects of generative AI, measures for data copyright and security protection are necessary.

Author

 

Toshiyuki Kita
Nextremer VP of Engineering

After graduating from the Graduate School of Science at Tohoku University in 2013, he joined Mitsui Knowledge Industry Co., Ltd. As an engineer in the SI and R&D departments, he was involved in time series forecasting, data analysis, and machine learning. Since 2017, he has been involved in system development for a wide range of industries and scales as a machine learning engineer at a group company of a major manufacturer. Since 2019, he has been in his current position as manager of the R&D department, responsible for the development of machine learning systems such as image recognition and dialogue systems.

Latest Articles

A summary of the latest examples of companies using generative AI! A thorough explanation of the benefits and implementation methods.

January 19, 2026

What is AI-based character recognition (AI-OCR)? A thorough explanation of the process, advantages, disadvantages, and key points to consider when implementing it!

January 19, 2026

A thorough explanation of how to collect data for machine learning! The steps to building a dataset and the benefits of outsourcing

January 16, 2026

How does object detection work? A thorough explanation of the amount of data required, use cases, and construction steps!

January 16, 2026