AI Projects and Workflow Sharing

On October 25, 2026, the 4th offline event of the Osaka AI Meetup was held in Osaka, themed "AI Projects and Workflow Sharing." Originally planned as a small gathering of about 15 people, the event ultimately attracted over 30 attendees, and the venue was changed to a larger space at short notice. The event focused on real-world practice rather than abstract concepts, sharing the processes, learnings, and unresolved questions from work actually completed recently. A total of five friends shared consecutively: 1. SOP for AI Short Drama Production: I kicked off the session, sharing the entire process of AI short drama production from idea to finished product. I focused on how to connect various production stages and how to refine a single attempt into a repeatable workflow. 2. How Experienced Programmers Use AI for Development: A-Du combined real cases such as document data extraction, business card design, and campus management systems to discuss what humans should do after delegating execution to AI: define the architecture, eliminate unsuitable features, and determine product direction. Even after a project is built, the question "should it be done this way" still needs to be answered. 3. Muse Usage and Experience: Qidi introduced everyone to Muse, leading to a discussion about general agents and personal AI applications—how to integrate new capabilities into daily tasks rather than just experimenting. 4. Harness Knowledge and Agent Working Environment: Teacher Wang spoke about the layer beyond the model: tools, context, permissions, task loops, and results validation. For agents to reliably complete tasks, business processes must first be clearly defined, data and tools must be connected, and there must be a way to verify completion. 5. Collaborative Development Project by Group Members: Jin Ge shared a project he is co-developing with several group members. Community interaction can go beyond sharing information—friends with similar interests and complementary experiences can truly push ideas forward together. The five topics ranged from content creation and software development to agents and multi-person collaboration. Although the perspectives differed, the conclusions were quite consistent: first find a real problem, then choose tools; clarify processes and judgments so AI can participate in execution more effectively; and trade-offs, acceptance, and cooperation always require human responsibility. After the event, a group of friends continued to chat over dinner, discussing everything from AI to entrepreneurship, technology, and life. Being able to hear real experiences from people in different fields up close, and bringing one's own questions into the discussion, is the very value of meeting offline~ Finally, I'd like to thank the five speakers for their diligent preparation, the over 30 attendees for their support, and the partners who helped arrange the venue. What the Osaka AI Meetup aims to do is build a continuously open AI exchange platform in Osaka. A project currently being explored, a process that has just become familiar, or even an attempt worth reviewing, can all be starting points for sharing. Friends in Osaka who are interested in AI, please feel free to join us, and we welcome you to share next time!

Photos from the day