I learnt to take the system's explanation back to the people doing the work

My first job after graduating from Nerikawa Institute was in a factory systems team at Hoshitora Instruments. I joined in April 1996 with an information engineering degree and a strong interest in how software represented physical work. I was one of ten analysts, without direct reports or personal budget authority. The experienced plant users supplied much of my practical education. When a record looked inconsistent, they could often explain a condition that our application did not capture. I learnt that a tidy data set could still give the next person an incomplete account of the factory.

I stayed at Hoshitora until August 2003. By the final period of that appointment, the company employed 1,800 people and recorded JPY38 billion revenue. Those figures describe the enterprise around my work, not a business or financial remit I personally controlled. My assignments covered interfaces, production information and support. I became increasingly interested in the arrangements that made a correction persist: who owned the definition, who could change it and how another shift learnt what had happened. A technical solution without those arrangements could return as the same operating problem a few months later.

I joined Kosemori Motion in September 2003 as plant systems manager. The role brought thirty-eight colleagues through four direct reports and, by its final period, JPY640 million annual authority. I was now responsible for continuing support as well as proposed changes. The company's 4,300 employees and JPY92 billion revenue at that point were larger than the business I had left, but scale was not the main adjustment. I had to make commitments to plant colleagues and then explain the consequences when a technical assumption failed. That experience made me more careful about distinguishing an installation date from an operating agreement.

In April 2010 I became head of production applications, leading seventy-two employees through five reports with JPY1.8 billion annual authority. The company had reached 5,900 employees and JPY158 billion revenue by the end of my appointment. Common applications offered opportunities to compare work across plants, but they also exposed differences in equipment, tooling and established practice. I began to treat an exception as a question to investigate rather than a defect in standardisation by definition. Some differences deserved removal; others needed a clear representation in the shared model. My team's job was to help the responsible operating owners distinguish them.

Enterprise integration became my remit in July 2015. I led 126 colleagues through five direct reports and held JPY4.6 billion annual authority by the final period. The work connected engineering information, enterprise planning and the records coming back from production. I learnt to ask whether two functions meant the same thing when they used an apparently common field. A plan, an authorised instruction and a completed operation could share an identifier without being interchangeable events. I wanted the enterprise to see that sequence clearly, because a mistaken equivalence could travel quickly through an integrated system and become difficult to challenge.

I became CIO in January 2021. Kosemori now makes gear reducers, couplings and control assemblies for industrial machinery at seven plants in Japan, Thailand and Vietnam. Its headquarters are in Nagoya. The listed enterprise has a 32% founding-family holding, with the remainder publicly held; 9,200 employees generated JPY286 billion revenue in the year ended March 2026. My own function has 198 employees, six direct reports and sixty-four contractors separately. FY2026–27 authority is JPY9.4 billion. I report to Reiko Tsumahiro and work closely with manufacturing director Masato Yurinobe, without taking over his responsibility for plant operations.

The decisions I discuss here concern the quality of the enterprise's explanation. We reconciled engineering and routing conflicts before allowing a migration result to stand as a reliable instruction. We limited an initial connectivity programme when twenty-one machine cells required further vendor and plant examination. We exposed infeasible orders in a planning trial and kept an image-inspection pilot in a supporting role after examining both its false flags and its misses. In each case, the technology result needed an operating interpretation. I could help produce the evidence and fund the response, but the other functions' participation was part of the decision rather than an approval to collect at the end.

I present to directors on technology investment and operating dependence, but I hold no board seat. I am considering how that experience might support a future external contribution in an established industrial enterprise. I would need to learn its particular manufacturing context and the broader responsibilities of a director, rather than assume that thirty years in systems leadership provides every answer. My proposed external capacity is eighteen days annually. The attraction for me is the chance to help another enterprise ask whether the factory described in its reports is sufficiently close to the one its people must run tomorrow.