Data Used
For this workshop, we used Zeolite’s actual revenue data — data accumulated in the data analytics infrastructure (DWH) that we have built through our ongoing support engagement. We connected this actual revenue data directly to the generative AI and pre-configured domain knowledge so that the AI could handle the data more appropriately.
Workshop Structure
The workshop was attended by 13 participants, including senior management, branch managers, and IT staff, with some joining remotely. Following an introduction by our team covering how the application works, how to use it, and key points to focus on in data analysis, participants worked on the theme of “Challenges in the revenue structure of your own department” through the following flow:
・Individual work
・Small group discussion
・Group presentations
Workshop Results (Participant Feedback)
In the post-workshop survey, over 80% of participants responded that they were “satisfied,” and likewise over 80% said they “want to use this in their work going forward.”
In particular, the ability to analyze data using everyday language received comments such as “it’s convenient because you don’t need to overthink it.” Anticipated use cases included profit and loss analysis, understanding revenue structure, and assisting in the preparation of materials for departmental meetings — scenarios directly tied to decision-making.
Most notably, a comment from one of the senior executives stood out: “It answered the questions I would normally ask staff in our department meetings.” We take this as a symbolic endorsement indicating that generative AI-powered data analysis is reaching a stage of practical readiness as a tool to support management decisions.
