NEWS

Press Releases
2025.08.21

MIXI Cuts Monthly Work Time by Approximately 17,600 Hours Through Company-Wide Use of ChatGPT Enterprise

~99% of users report improved productivity, with approximately 11 hours saved per person each month~

Tokyo, Japan, August 21, 2025 — MIXI, Inc. hereby announces the results achieved through the use of ChatGPT Enterprise, OpenAI’s enterprise generative AI service, which was introduced to all employees*¹ in March 2025.

Following the company-wide rollout of ChatGPT Enterprise, we have been working to create an environment where all employees can effectively use it in their work. Specifically, together with OpenAI, we have provided internal training programs to support the independent use of generative AI and foster advanced AI talent, including ChatGPT 101 Training for all employees, workshops for new graduate hires joining in 2025, and a hackathon for engineers focused on product development using the OpenAI Agents SDK.
These initiatives have helped a corporate culture of using generative AI effectively steadily take root. In less than three months after the company-wide rollout, ChatGPT Enterprise’s active user rate (WAU) reached 80%, and more than 1,800 custom GPTs*² were created voluntarily. In addition, a wide range of use cases has emerged across departments, and some projects have delivered remarkable results, including reducing work time by more than 90%. Below are the main results confirmed three months after the company-wide rollout.

Three-Month Highlights
The number of custom GPTs has roughly tripled, with an estimated monthly reduction of about 11 work hours per person.

Training programs have helped spread practical knowledge of generative AI use across teams, and the number of custom GPTs has grown to about three times its level at launch. Tools for day-to-day work support that were once built by engineers on request are now increasingly being created by employees themselves, and more employees are taking the initiative in process improvement. Some non-engineering employees have even said it makes them feel like programmers, showing that their day-to-day thinking and ways of working are beginning to change. ChatGPT Enterprise also continues to see high usage, with MAU exceeding 90% and WAU reaching 80%.

Three months after the company-wide rollout of ChatGPT Enterprise, a usage survey*³ found that it is expected to reduce work time by approximately 17,600 hours*⁴ per month across the company. This is equivalent to a reduction of about 11 hours per user per month. In addition, 99% of users felt their productivity had improved, and 89% felt more satisfied with their work.

・Time savings: 90% said they were able to save time. Of those, 22% saved 5-10 hours per week or more.
・Productivity improvement: 99% felt their productivity had improved. Of those, more than 80% said it had improved to at least a moderate degree.
・Contribution to creativity: 67% felt that their creativity improved. Of those 20% strongly agreed.
・Job satisfaction: 89% felt that their job satisfaction had improved. Of those, 26% strongly agreed.

When asked about the main ways they use it in their day-to-day work, the most common response was creating and editing written content. This was followed, in order, by brainstorming and developing creative ideas, research such as online research, data searches, and fact-checking, technical support or troubleshooting such as diagnosing and resolving technical issues, and data analysis and generating insights.

Use cases leveraging ChatGPT Enterprise are reported monthly by each department, and frontline-led AI adoption is spreading company-wide.

ChatGPT Enterprise Use Cases Across Departments

・Compliance Department, Legal Affairs Division: Introduction of a custom GPT for Terms of Service review
Reviewing terms of service, which used to take 30 minutes to an hour, can now be completed in about 10 minutes, saving approximately 40 work hours per month. It also reduced workload and helped cut outsourcing costs.

・FamilyAlbum Business Department, FamilyAlbum Marketing Division: Using GCT: Creative Planning to Improve Planning Efficiency
Reduced monthly planning workload by approximately 28 hours for FamilyAlbum, the international version of FamilyAlbum. It also eased review workloads and enhanced the team’s proposal capabilities, contributing to overall productivity and creativity.

・Corporate Promotion Department, FP&A Intelligence Division: Building an Assistant Bot to Streamline Budget-Related Applications and Inquiries
Rolled out the bot company-wide to handle budget-to-actual questions and reduce the workload for applicants. Bot support simplified the application process for the project management tool, cutting the time required per request by about 50%. This is expected to improve efficiency by about 150 hours annually.

・Corporate Development Department, Investment Division: Hypothesis validation using “VC Fund Initial Review Supporter”
Automates information organization and report creation for the initial review of startups and VC fund investments. Cuts work time by up to 90% per case, equivalent to an annual reduction of approximately 108 hours.

This initiative is also available below, published by OpenAI as a case study on its implementation in Japan.
https://openai.com/index/mixi/

Going forward, we will further expand the organization-wide knowledge base and leverage digital programs provided by OpenAI to drive broader adoption and more advanced use in our day-to-day operations.

Highlights from the 2025 Workshop for New Graduate Employees

Scenes from the Internal Hackathon for Engineers

■ Comment from Tatsuma Murase, Director and Senior Corporate Officer, MIXI, Inc.

Array

The company-wide adoption of ChatGPT Enterprise is not just about improving efficiency. It is about building the foundation for an environment where every employee can focus on more creative work. The fact that more than 1,800 custom GPTs emerged organically across the company within just a few months of rollout is proof that AI is beginning to take root not as a tool people are forced to use, but as a capability they can truly master. More than 300 AI-related projects are currently underway across the company, many of which have already been implemented in day-to-day operations. Given the speed of change driven by generative AI, how quickly we can move into the stage of truly mastering it will have a major impact on our future business competitiveness. We see this as the right moment to take a proactive approach and update both the way we work and the way we create value with AI at the core.

■ Our AI Initiatives

Since 2023, we have promoted the company-wide use of generative AI. We have proactively introduced and operated advanced AI technologies across the company and within individual divisions, and have continuously pursued initiatives to improve operational efficiency. In December 2024, the AI Committee, overseen by Director Tatsuma Murase, was launched, further accelerating AI adoption across the company. Based on our internal guidelines, we are now fully integrating various generative AI tools into business operations while strictly ensuring appropriate information input in accordance with confidentiality labels. In addition, we have assigned AI ambassadors to each division to drive AI adoption from the ground up. Going beyond efficiency gains at the individual level, more than 300 AI-related projects were underway as of the end of July 2025, focused on cost optimization, value creation, and new business planning. In addition to ChatGPT Enterprise, we also use multiple generative AI tools, and the AI usage rate among all employees has reached 99%.

Going forward, we will continue to establish generative AI as part of our business infrastructure while fostering an environment where employees can fully exercise their creativity. Through AI, we aim to create new experiential value that expands connections between people, while continuing to explore the possibilities of MIXI’s unique blend of communication and AI.

*1) Applies to all employees with a company-domain email address (full-time employees, contract employees, expert employees, part-time employees, temporary staff, and independent contractors), approximately 2,000 as of April 2025.
*2) A feature that lets you customize ChatGPT to suit your needs and create an original chatbot (AI assistant).
*3) Valid responses: 338 cases
*4) Calculation method: Responses to the question, “On average, how many hours per week does ChatGPT help save on work-related tasks?” (n=329) were aggregated using the lower bound of each option, resulting in 836 hours/week. This was multiplied by the average number of weeks per month, 4.33, to calculate 3,620 hours/month. Converted to an active user rate of 80% (1,600 people), this comes to approximately 17,600 hours/month.

  • Home
  • News TOP
  • MIXI Cuts Monthly Work Time by Approximately 17,600 Hours Through Company-Wide Use of ChatGPT Enterprise