In-house AI Development Support Through a Two-Month Training Program

Outburn Inc. (Headquarters: Chuo-ku, Tokyo; CEO: Kodai Moriyama), which supports organizational AI utilization with the mission to "Contribute to the change of people and organizations," has conducted a business improvement workshop and practical generative AI utilization training for 15 cross-departmental members of Global Produce Inc. (Headquarters: Shibuya-ku, Tokyo; President: Masaki Mitsuhata). This training was provided to 15 selected cross-departmental members, totaling 6 sessions over 18 hours, and provided consistent support from work inventory to practical generative AI application, and prototyping and verification of AI tools tailored to actual business practices. As a result of the training, participants' business improvement themes have become more concrete. One example is the prototyping of an AI tool to support the judgment of rejections in expense processing. Moving forward, efforts are planned for internal deployment with the goal of reducing rejections by 50%. Background of Training Implementation While interest in generative AI utilization in companies is increasing, many on-site challenges are observed, such as "not knowing where to start" and "not seeing how to establish AI adoption in actual business practices." In light of these circumstances, this training emphasized not just learning how to operate tools, but having each participant organize their own tasks, clarify which processes AI can be utilized in, and then translate this into concrete improvement themes. Implementation Overview Training Name Business Improvement Workshop / Practical Generative AI Utilization Training Implementation Time Total 6 sessions, 18 hours (6 hours for Business Improvement Workshop + 12 hours for Practical Generative AI Utilization Training) Number of Participants 15 people (cross-departmental selected members) Implementation Format On-site support (Including workshops, chat support, and Q&A sessions) Main Tools Used ChatGPT, Genspark, Perplexity Features of this Training This training did not prioritize explaining AI tool functions first. Instead, it adopted an approach of first organizing the current situation through work inventory and clarifying the business processes eligible for AI utilization. Furthermore, by focusing on hands-on sessions where participants actually work with the tools rather than just receiving knowledge input, the design was made to facilitate the immediate connection of understanding to practical application. Additionally, by setting themes and prototyping tailored to each participant's specific work content, support was provided with an awareness of on-site adoption that extends beyond the training period. Outcomes Achieved Through Training Through this training, participants visualized their daily tasks and concretely organized areas for AI application, leading to the planning and prototyping of improvement measures directly linked to actual business practices. Specifically, in accounting operations, an AI tool was prototyped to support the judgment of rejections during expense processing, concretizing a business improvement measure targeting a 50% reduction in rejection cases. In addition, cases were confirmed where some tasks that previously took about 30 minutes of manual work could be shortened to about 2 minutes through AI utilization. After the implementation, a change in perception was observed, where AI is viewed not as something that "completely replaces tasks," but as "something that supports and enhances parts of the business," leading to the fostering of practical understanding in the workplace.