What I want directors to understand before they accept my recommendation

An enterprise system can make an operating proposition travel farther than its original authors expect. A planning output may become a customer-date discussion, an engineering record may become a plant instruction and a model score may become an investment claim. When I present to directors, I try to explain which of those transitions the evidence supports and where another decision remains necessary. The board should not need to inspect every application field. It should be able to understand whether management has identified the authority, conditions and continuing obligations behind the recommendation.

I have presented as Kosemori's CIO since January 2021. I am not a director. My executive responsibility includes 198 employees and JPY9.4 billion authority for FY2026–27, while the company itself has 9,200 employees and JPY286 billion revenue for the year ended March 2026. Keeping those scopes distinct is part of the same discipline. A large enterprise does not make every executive responsible for all of its decisions, and a technology budget does not confer manufacturing or quality authority. I explain the working relationship with those functions as part of the proposal rather than list them only as people who will be consulted later.

In the work-order migration, eighty-four disagreements needed engineering or production resolution before the examined records could be treated as agreed instructions. I presented the three-week phased introduction and JPY430 million cost alongside that requirement. The board-level issue was whether the route to agreement was credible and proportionate. It was not an invitation for directors to decide individual revisions. Once the conflicts were reconciled, I kept the scope of that result clear: the examined orders had been addressed, while a continuing process was still needed for later changes.

The inspection pilot required another kind of explanation. A held-out set of 18,000 images produced 360 correctly flagged defects, seventy-two false flags, twenty-eight missed labelled defects and 17,540 correctly unflagged images. The counts described a controlled evaluation, not actual production escapes. I supported a decision-support use at JPY110 million pilot cost, with quality authority retained. Directors could then examine the proposed next step, the evidence still required and the continuing ownership, rather than be asked to approve an ambitious use on the strength of one favourable score.

I also distinguish technical measurement from an operating benefit. Aligning readings for forty-eight machines made the agreed period more usable for analysis. It did not itself reduce consumption. The JPY85 million programme needed an account of why better measurement was worthwhile and what subsequent decisions it could support, without borrowing savings from an intervention that had not happened. Product mix and utilisation remained relevant to comparison. I want those qualifications to survive when the programme's figures move into a board pack or a shorter executive summary.

A future external seat would require a broader contribution than these presentations. I would need to join fellow directors in assessing the enterprise's overall position and understand matters beyond my technology experience. I would seek induction in the company's finances, operating context and governance responsibilities, and would examine appointment-specific conflicts before accepting. Eighteen proposed annual days would need to fit the actual preparation, committee and exceptional demands. I would not infer independence from having no current directorship or use my executive title as evidence of competence for every committee.

The practice I could contribute is to ask where an apparently settled account still depends on a decision somebody else must own. That question can improve oversight without drawing directors into day-to-day management. It encourages a report that preserves both the achieved result and its boundary, allowing the next decision to be made on a clearer basis. I would want management to feel able to explain an exception or a limited trial honestly, while expecting directors to challenge the cost, relevance and proposed response. That combination is more useful than either unquestioning confidence in a system or general suspicion of technical change.