Magconn Inc. conducted a generative AI training program called "Corporate Reskilling" for six participants over approximately seven months, from February to August 2026. Self-assessments of Copilot utilization skills increased from an average of 1.60 to 3.00, a rise of approximately 881 TP3T, and the frequency of AI interaction increased from 4.0 days per week to 5.0 days per week. All participants now use AI in their work every business day.

At the start of the training, three out of the six participants "didn't know how to operate the equipment and had almost no experience using it." This article describes the six months it took for this company, which had previously changed the usability of its equipment with magnetic connectors, to now change its own work methods.

Company Introduction

Magconn Corporation, established in 2019 and headquartered in Kokubunji City, Tokyo, is a manufacturer of magnetic connectors. They handle everything from development to manufacturing and sales of magnetic connector modules, tablet cases, charging stands, and related peripherals, with major restaurant chains and electronics manufacturers nationwide as their main clients.

At the heart of the company is its unique connector that uses magnetic force for attraction."Magconn"This technology, which enables charging and data synchronization without the need to plug and unplug cords, has been patented both domestically and internationally, and utilizes a contact charging method.Charging efficiency 97%This is achieved through features such as spark and short-circuit prevention, corrosion resistance due to gold plating on the terminals, and waterproof, splashproof, and dustproof specifications, all designed with the harsh environments of professional use in mind.

Our own branded product, which packages this technology, is a business solution for tablets."Pitatto-kun"We address the inconveniences faced by users of tablet devices, which are essential for self-ordering systems, such as their weight and difficulty in disinfection and cleaning, by designing cases and stands. Through our three businesses—module sales, OEM/ODM development and manufacturing, and planning, manufacturing, and sales of our own branded products—we have expanded into a wide range of fields, including shopping carts and medical equipment.

The company name "Magconn" is a portmanteau of "Magnet" and "Connectivity." The approach of resolving the inconveniences inherent in the simple act of plugging and unplugging by changing the way connections are made aligns with our current approach to utilizing generational AI.

Issues before introducing generative AI

The training program began with the CEO's own desire to eliminate waste in business operations. Within a small team handling everything from development to manufacturing and sales, the starting point was how to lighten the workload for each individual.

On the other hand, there were differences in the participants' status at the start of the program. In their self-assessment of Copilot usage skills, 3 out of 6 participants were at level 1, meaning they "did not know how to operate it and had almost no experience using it." On average, they interacted with AI 4.0 days a week, and each person used AI for an average of 2.17 tasks. Although they interacted with AI for most of the week, its use was limited to tasks such as translation and drafting, and it had not yet reached the core of their work.

At the beginning of the training, each of the six participants took stock of their assigned tasks. They identified 27 tasks, totaling 8,220 minutes (137 hours) per week. Looking at the breakdown, there was a clear imbalance in how time was being used.

Time required for each task category (top 5)

Business Category	Time required per week	Percentage of the whole
Document creation	2,310 minutes (38.5 hours)	28.1%
Customer support	1,050 minutes (17.5 hours)	12.8%
Email correspondence	750 minutes (12.5 hours)	9.1%
Strategic Planning	600 minutes (10.0 hours)	7.3%
Customer development	600 minutes (10.0 hours)	7.3%

The top five categories alone account for approximately 651 TP3T of the total. In particular, document creation, specifically proposals and quotations, was the biggest burden shared by all three sales representatives, yet at the time of inventory, all of them responded that they were "not using AI." The heaviest tasks remained untouched.

Implementation details - Magconn's unique initiatives

The six-month program was designed around the two main goals of aligning everyone's efforts within a small organization and directly linking the program to the most pressing tasks identified during the inventory review.

• All six participants attended.
  The company president himself participated in all sessions as one of the trainees. The aim is to avoid a situation where "only certain people can use the information" by having people from all positions and job roles learn the same content at the same time.
• The series consists of nine episodes, with the halfway point being an interim report.
  The program consisted of nine sessions in total: one startup meeting, four workshops, one interim report meeting, and three implementation support meetings. Video learning was conducted concurrently throughout the entire period, designed to allow for a back-and-forth between learning and practice.
• Work design starting with inventory
  The workshop topics were selected from the top-performing tasks among the 27 identified tasks: document creation, email correspondence, and data aggregation. Rather than being general exercises, the participants were encouraged to adapt the procedures so they could apply them to their own work starting the very next day.
• Practical applications focused on Microsoft 365 Copilot
  We focused on using AI within applications that everyone already uses daily, such as Outlook, Excel, PowerPoint, and Teams. By not adding new tools, we shortened the time it took for users to adopt the new technology.
• We've reached the starting point of business process automation.
  In the final support meetings, we covered Office Scripts and Power Automate. The content involved recording and automating routine Excel processes and connecting them to monthly summary notifications, and the completed code was distributed to everyone.

Benefits of using generative AI

The biggest change was in the participants' own skill perception. Three participants who were at Level 1, meaning they "didn't know how to operate the equipment," all moved up to Level 3 or higher through the training, leaving zero participants at Level 1. One sales representative reached Level 4, experiencing tangible results across multiple tasks.

Changes in key indicators (average for participants)

index	change	stretch
Copilot utilization skills	1.60 → 3.00	Approximately 88% increase
Time saved per week	3.92 hours → 5.10 hours	Approximately 30% increase
Number of jobs using AI	2.17 items → 3.20 items	Approximately 48% increase

The time saved by AI increased from 3.92 hours per person per week to 5.10 hours, and the number of tasks utilizing AI increased from 2.17 to 3.20 per person. The frequency also increased from 4.0 days per week to 5.0 days per week, with all respondents in the final survey answering "5 days a week." This indicates that AI is not being used all at once on specific days, but rather is being integrated into daily tasks.

* The figures are self-reported values from participant questionnaires. Skills, time saved, and number of tasks utilized are averages before and after implementation, and the rate of change is calculated as (after implementation - before implementation) ÷ before implementation. The time required for each task category is the value reported by the participant in the task inventory sheet.

The widespread use of the training is also reflected in the words of the participants. Participants who had no prior knowledge at the start of the training gave the following feedback:

"I had absolutely no prior knowledge, but through training I learned how to incorporate it into my work, and I've gradually been able to do so."

From participant surveys

One of the reasons the exercises became so widely adopted was that the topics were tailored to the company's actual work.

"By conducting workshops that included practical usage methods and examples, we've been able to utilize the technology in our actual work."

From participant surveys

Furthermore, some employees have reported that the very structure of their work has changed. For example, they now process data received from shipping companies, feed it into AI, and create shipping report emails for each client all at once—combining the procedures they learned in training and completely redesigning the workflow for their assigned tasks.

"We process and import data from shipping companies into our shipping notification emails, and then have them create emails for each client all at once."

From participant surveys

There was an example of a student who was struggling to come up with prompts, but then realized that "the AI can even come up with the prompts in the first place," and switched to proceeding while consulting with the AI. We are moving from the stage of learning how to operate the system to the stage of designing procedures with AI in mind.

Future outlook

With all six participants having reached Level 3 or higher, Magconn's next theme will shift from "being able to use it" to "being able to continue using it and achieve results."

• Putting the distributed automation code into production.
  The code created with Office Scripts and Power Automate has been distributed to everyone. The next step is to implement it in actual operation, starting with recurring processes such as monthly summary notifications.
• Full-scale application to document creation
  The most demanding task, accounting for 2,310 minutes per week, is document creation, primarily consisting of proposals and quotations. Several respondents in the survey cited "creating slides for proposals and reports" as a task they would like to master, making it the area with the greatest potential for growth.
• The establishment of data aggregation and graphing.
  The most frequently cited task as "the most time-consuming and laborious task" among participants was data aggregation, graphing, and analysis. Deepening your use of Copilot in Excel will directly reduce the time spent creating monthly reports.
• Creating internal usage rules
  Given the nature of our business, which involves handling customer design information through OEM/ODM, we want to avoid a situation where the decision of what to input into the AI is left to individual discretion. Now that everyone is using it on a daily basis, it's time to establish common rules.

Through this six-month initiative, Magconn transformed from a company where "some employees didn't know how to use AI" to a company where "everyone uses AI every business day."

Having addressed the inconveniences inherent in the simple act of plugging and unplugging by re-examining the connection methods, the company, whose name is a combination of Magnet and Connectivity, next focused on improving the connections within their own work.