Pico Systems Co., Ltd. conducted a generative AI training program called "Corporate Reskilling" for 11 employees over a six-month period from January to June 2026. According to self-reported surveys by the participants, the average number of tasks utilizing AI per person increased from 2.2 at the start of the training to 3.8, and the average weekly time saved increased from 2.0 hours to approximately 3.0 hours.

At the training results presentation, various initiatives were shared to automate tasks that previously relied on individual experience, such as initial contract checks and training new system engineers, using AI. From "learning" to "using in the organization"—we introduce the journey of this company, based in Okayama, which has supported corporate systems for over half a century.

Company Introduction

Pico Systems Co., Ltd. is a systems company headquartered in Okayama City. It began selling office equipment in 1968 and started handling computers in 1971. The company changed its name to Pico Systems Co., Ltd. in 1975. In 2026, it celebrated its 59th anniversary. In addition to its development center in Okayama, it has offices in Tokyo, Osaka, and Himeji, and employs 161 people (as of May 2026).

We handle a wide range of corporate IT environments, from custom system development and support for implementing manufacturer packages to developing original business packages, building cloud systems, selling and maintaining office automation equipment, and even operating docomo shops. Our motto is:"CREATE THE FUTURE"is.

Eleven people from the administrative, sales, SE, and service departments participated in this training.

Issues before introducing generative AI

When introducing the training program, the company focused on members of the administrative department. The administrative department had a lot of paper documents, and manual review was time-consuming. In sales, the time it took to create proposals was a challenge. As a solution, they decided to introduce a generation AI.

As a systems company, we have always strived to eliminate manual work by implementing systems. However, we had not yet established a path for each employee to use AI in their own work. At the start of the training, we received the following feedback:

Since creating documents often takes a lot of time, I want to learn effective prompts to reduce the time spent.

From participant surveys

At the first workshop, held shortly after the start of the training, each participant was using AI for an average of 2.2 tasks, and the time saved by AI was 2.0 hours per week. Some participants started with almost no use of AI, while others were skeptical about using AI.

Implementation Details - A unique initiative by Pico Systems.

• A total of four workshops, progressing in stages according to each task.
  In the first session, we covered the basics of AI, text generation, and data visualization; in the second session, document analysis such as contracts and meeting minute generation; in the third session, image generation and idea generation; and in the fourth session, chatbot creation and knowledge creation of company regulations. With each session, the topics progress from individual tasks to systems used within an organization.
• Identify uses through a business inventory review.
  In addition to the survey, we inventoryed the tasks of 11 individuals and organized them into 79 tasks. From these, we identified 54 tasks where AI is already being used and 48 tasks where AI could be used in the future (in some cases, multiple uses can be identified for a single task), which helped us identify which tasks are large in scale and therefore easy to use with AI.
• Connecting learning to practical work through after-sales support
  We reviewed all four sessions, checked each participant's application progress, and helped each individual identify themes to try in their actual work. The themes decided here are then used as topics for subsequent support meetings.
• Share your efforts at the results presentation meeting
  The second support meeting for utilization served as a results presentation session, where participants presented slides summarizing their work-related initiatives. Following the presentations, each presenter received feedback outlining their strengths and areas for improvement.

Benefits of using generative AI

What particularly caught the eye at the results presentation was the initiative to incorporate AI into tasks that previously relied on individual experience. At the presentation, a system was presented that uses AI to perform an initial check of outsourcing contracts and non-disclosure agreements, as well as a chatbot that answers questions based on company regulations. The chatbot has been shared within the company and is available for anyone to use.

We also looked at the data from the participant questionnaires (self-reported). We compared the responses from the first workshop, held immediately after the start of the training, with the most recent responses from each participant, using the same 11 individuals.

Changes in key KPIs (11 participants)

The average satisfaction level among the respondents (9-11 people) in each session ranged from 4.27 to 4.56 on a 5-point scale.

Initiatives shared at the results presentation

In training new system engineers, we've created a system where AI plays the role of a customer and evaluates each stage, from initial consultation to requirements gathering and design. This initiative aims to standardize the review and training process, which previously varied depending on the individual, by using the same criteria.

At the results presentation, they introduced an initiative where, when a customer consults with them, they organize the conditions such as function, scale, budget, and security, provide them to the AI, and then narrow down the candidate products with a comparison table. Because the candidates can be narrowed down before final confirmation with the manufacturer, the preparation of proposals can be expedited. Those who have started using AI for the majority of their sales and planning work also shared their personal experience of reduced overtime.

The participants' comments reflect a shift in how they interact with AI.

I was a little skeptical about using AI, but after trying it, I found it to be very convenient and efficient. While it's useful when used correctly, it also makes many mistakes, and in some cases, it can actually decrease efficiency. I think that its accuracy will improve with continued use, so I hope we can use it throughout the entire company.

From participant comments

These users are saying they understand when to use it, not only for its convenience but also to be mindful of making mistakes. Some have even started incorporating it into their daily work, going from hardly using it at all.

We started with almost no experience using AI, but I'm glad that we were able to gradually use AI in our daily work outside of the training sessions as well. I feel that it's leading to increased efficiency, albeit gradually, based on what we learned in the training. The instruction templates were particularly helpful.

From participant comments

Furthermore, the idea of separating the tasks that should be handled by AI from those that require human judgment has emerged.

I've come to realize once again that AI isn't a magic tool, and that the results depend heavily on the user's mindset and the quality of their questions. After the training, I immediately began to organize the parts of our work processes that can be handled by AI and the parts that require human judgment, and I'm gradually putting this into practice.

From participant comments

Those who spearheaded the training reflect that, in a company that had previously eliminated manual work by implementing systems, this initiative has paved the way for employees to proactively use AI to transform their work processes.

Future outlook

Now that we're starting to see results, the next point of discussion is how to establish the announced initiatives as part of the organization's systems.

• Reduce manual work in administrative departments by 20-30%
  Those who spearheaded the training have set a goal of reducing manual work in the administrative department by 20-30%. They also emphasize that it's necessary to proceed with proper management rather than having each person create systems independently.
• Turn your initiatives into templates and chatbots
  By organizing the prompts and systems shared at the results presentation into templates and chatbots for each task, we can ensure that employees who did not participate in the training can use them in the same way. The 48 potential applications identified during the task inventory will serve as the basis for this.
• Using the right tools for your job
  The tools best suited to different job roles, such as document creation, data analysis, and development, vary. Organizing how to use different tools depending on the department and job role is another topic to consider.

Through a six-month program, Pico Systems Co., Ltd. has moved from the stage of "learning" generative AI to the stage of "using it within the organization." One of the results achieved during this process is that the average number of tasks utilizing AI per person, as reported by the participants, increased from 2.2 to 3.8, and numerous initiatives emerged to turn individual skills into systems.

"CREATE THE FUTURE"—With over half a century of supporting corporate systems, the company is entering its next phase, now equipped with a new tool: generative AI.