Manufacturing Company A (Major Manufacturer)

The Challenge

With every new product launch or design change, the team had to rethink processes from scratch even when similar projects existed, leading to long lead times. In addition, knowledge such as manufacturing standards and past incident records was scattered across multiple locations, making retrieval heavily dependent on individual experience and memory. This resulted in increased review workload, rework, and over-reliance on specific individuals.

Why Stella Maris

What the company needed was not simply a system to manage fixed standard flows, but a mechanism to quickly assemble the necessary processes based on the unique conditions of each project. They therefore explored an approach using generative AI—leveraging past standards, incident records, and existing processes—to support the creation and review of process drafts.

Our Approach

The team explored a system in which AI references related documents and past cases based on natural-language instructions to automatically generate role-specific process drafts. Responses and generated content were linked to the source documents and rationale, enabling users to verify the basis for decisions. A dashboard concept was also designed to provide an overview of process improvement status across departments and production lines, aiming to balance individual process improvement with overall optimization.

Results

Through the proof of concept, a clear direction emerged for streamlining the per-project process reviews that had previously occurred every time, by leveraging accumulated knowledge. In particular, the ability to quickly generate initial drafts informed by relevant incident records and standards was confirmed to have the potential to accelerate the start of reviews, reduce rework, and improve process accuracy. A concrete vision was also developed for codifying the judgment rationale of experienced staff, contributing to junior employee development and strengthening quality governance.

Solutions Used