Revolutionizing Cabin Crew Workflows: Fujitsu and Headwaters Pioneer On-Device Generative AI for Japan Airlines
In a significant advancement, Fujitsu Limited and Headwaters Co., Ltd., a leading AI solutions provider, have successfully completed field trials of a groundbreaking generative AI solution. This innovative technology is engineered to optimize and enhance the efficiency of handover report creation for cabin crew members at Japan Airlines Co., Ltd. (JAL). The trials, conducted from January 27 to March 26, 2025, have demonstrated substantial time savings and hold the promise of transforming how cabin crews manage their operational workflows. This collaboration marks a pivotal moment in leveraging AI to improve operational efficiency within the aviation sector.
The Challenge of Handover Reports
Currently, JAL cabin crew members dedicate a considerable amount of time and effort to the creation of detailed handover reports. These reports are indispensable for ensuring a smooth and effective transfer of information between incoming and outgoing cabin crews, as well as ground staff. The reports meticulously document passenger needs, potential safety concerns, equipment status, and any other pertinent information necessary for clear and comprehensive communication. This thoroughness, while vital for operational efficiency and guaranteeing passenger safety, can be a laborious and time-intensive task. This diverts valuable time and attention away from direct passenger interaction and other critical responsibilities that are essential to the overall flight experience.
The comprehensive nature of these reports requires cabin crew to meticulously document every relevant detail of the flight. This includes everything from special meal requests and medical conditions of passengers to any malfunctions or irregularities observed with the aircraft’s equipment. The process also entails gathering and consolidating information from various sources, such as passenger manifests, maintenance logs, and pre-flight briefings. The preparation of these reports often extends beyond the duration of the flight itself, requiring cabin crew to dedicate additional time after landing to finalize and submit the documents.
Fujitsu and Headwaters recognized this challenge as a significant pain point for cabin crew members and sought to develop a solution that would alleviate the burden while upholding the accuracy and completeness of the handover reports. Their solution capitalizes on the power of generative AI to automate and streamline the report creation process. This enables cabin crews to focus more intently on delivering exceptional service and prioritizing passenger well-being throughout the flight. By automating the more mundane aspects of report generation, the AI solution frees up cabin crew to address more pressing needs and provide a higher level of personalized attention to passengers.
The Innovative AI Solution
To directly address the issue of time-consuming report generation, Fujitsu and Headwaters strategically utilized the capabilities of Microsoft’s Phi-4, a compact yet powerful language model (SLM) specifically optimized for operation in offline environments. This strategic decision was paramount in circumventing the need for a large language model (LLM), which would necessitate constant cloud connectivity. Instead, they developed a user-friendly, chat-based system accessible on standard tablet devices. This empowers cabin crews to generate comprehensive reports efficiently both during and after flights, regardless of the availability of a stable internet connection.
The on-device nature of the solution offers several indispensable advantages that are particularly relevant to the operational demands of the aviation industry. First and foremost, it ensures the highest levels of data privacy and security, as sensitive passenger information and operational details remain securely confined within the device itself and do not require transmission over a potentially vulnerable network. This is particularly crucial within the aviation industry, where data security and compliance with stringent privacy regulations are of paramount importance. The ability to process and store data locally eliminates the risk of unauthorized access or interception of sensitive information during transmission.
Secondly, the offline functionality guarantees unwavering availability of the solution, irrespective of the presence or strength of internet connectivity. This is essential for cabin crews who frequently operate in remote areas or traverse airspace with limited or no network coverage. The AI solution remains fully operational, enabling them to generate and update reports without interruption, even in the absence of a reliable internet connection. This ensures that critical information is always accessible and that handover reports can be completed promptly and accurately, regardless of the flight’s location or circumstances.
Finally, the utilization of a small language model (SLM) plays a critical role in minimizing the computational resources required to run the AI effectively. This makes the solution exceptionally well-suited for deployment on standard tablet devices, which typically have limitations in processing power and battery life. By optimizing the AI model for efficiency, Fujitsu and Headwaters have ensured that it can operate seamlessly on readily available hardware without impacting the performance or battery life of the tablet devices. This ensures that cabin crew members can rely on the AI solution throughout the duration of their flights without experiencing any technical difficulties or limitations.
The comprehensive trials conducted by Fujitsu and Headwaters have conclusively demonstrated that the solution not only significantly assists in the creation of high-quality, accurate reports but also achieves a substantial reduction in the amount of time spent on report generation. This reduction translates directly into increased efficiency for cabin crew members. This allows them to dedicate a greater proportion of their time and attention to passenger care, addressing passenger inquiries, and performing other essential tasks that contribute to a positive and seamless travel experience. The time savings achieved through the AI solution enable cabin crews to be more proactive in anticipating passenger needs, providing personalized assistance, and ensuring the overall comfort and well-being of everyone on board.
Roles and Responsibilities
The successful implementation of this innovative solution was a direct result of a highly collaborative effort between Fujitsu and Headwaters, with each company contributing its unique and specialized expertise and resources to the project.
Fujitsu, leveraging its advanced Fujitsu Kozuchi AI service, assumed the critical responsibility of fine-tuning Microsoft Phi-4 through the utilization of JAL’s extensive repository of historical report data. This meticulous process involved carefully training the AI model to accurately understand the specific language, terminology, and formatting conventions consistently used in JAL’s handover reports. By tailoring the AI model precisely to the unique needs and operational requirements of JAL’s cabin crew, Fujitsu ensured that the generated reports would be highly accurate, relevant, and easily understandable for all stakeholders. The fine-tuning process involved feeding the AI model with a vast amount of historical data and iteratively refining its algorithms until it could reliably generate reports that met JAL’s stringent quality standards.
Headwaters, on the other hand, focused its expertise on developing a robust and highly effective business-specific generative AI application utilizing Phi-4 as its core engine. This involved the strategic utilization of advanced quantization technology to optimize the AI model for efficient performance on standard tablet devices operating in offline environments. Headwaters’ team of experienced AI consultants played a crucial role in thoroughly analyzing existing workflows, identifying key areas for improvement, and designing an intuitive and user-friendly interface for the chat-based system. They also provided expert guidance on AI implementation best practices, conducted comprehensive evaluations of the trial results, and managed the agile development process with meticulous attention to detail. In addition, Headwaters’ AI engineers skillfully built a fine-tuning environment for Fujitsu Kozuchi and provided comprehensive technical support for optimization tailored precisely to JAL’s specific usage environment and operational needs.
Expert Perspectives
Shinichi Miyata, Head of Cross-Industry Solutions Business Unit, Global Solutions Business Group, Fujitsu Limited
Shinichi Miyata conveyed his enthusiasm and satisfaction regarding the successful application of generative AI within Japan Airlines’ cabin operations. He emphasized that the joint proof-of-concept represents a significant leap forward in advancing the practical utilization of generative AI in demanding offline environments. Miyata highlighted the transformative potential of this technology to revolutionize operations across a wide spectrum of industries and roles where consistent and reliable network access is often limited or entirely unavailable. He attributed the overall success of the collaborative effort to the exceptional proposal capabilities of Headwaters, combined seamlessly with Fujitsu’s deep technological expertise in the field of artificial intelligence. Looking forward, Miyata reiterated Fujitsu’s unwavering commitment to strengthening its strategic partnership with Headwaters to effectively support customers’ business expansion and address pressing societal challenges through innovative AI solutions.
Yosuke Shinoda, CEO, Headwaters Co., Ltd.
Yosuke Shinoda expressed his deep honor and privilege in being a part of the pioneering initiative to demonstrate the profound efficiency of utilizing generative AI for cabin crew reporting, working alongside Fujitsu and Japan Airlines (JAL). He underscored the technology’s remarkable ability to substantially reduce the time necessary to both prepare and meticulously correct handover reports, highlighting its immense promise for continued and expanded application across various operational areas. Shinoda extended his sincere gratitude to Microsoft Japan for its outstanding technical capabilities and unwavering support throughout the project. He expressed his eager anticipation for continuing to work closely with Fujitsu to support JAL in the practical and widespread application of generative AI to further enhance its operations.
Keisuke Suzuki, Executive Officer, Head of Digital Technology Division, Japan Airlines Co., Ltd.
Keisuke Suzuki expressed his sincere delight in conducting the proof-of-concept for a cutting-edge generative AI solution specifically aimed at significantly enhancing the efficiency of cabin crew operations through the collaborative partnership with Fujitsu and Headwaters. He emphasized that by strategically leveraging the power of generative AI, JAL aims to streamline the often-complex handover report creation process and substantially reduce the burden on its valued cabin crew members. This reduction allows cabin crew members to dedicate more of their time and attention to providing more personalized and attentive service to each and every customer. Suzuki expressed his anticipation for further and continuous improvement of customer service through the continued development and implementation of this transformative initiative.
Tadashi Okazaki, Corporate Executive Officer, Senior Managing Director, Cloud & AI Solutions Business Division, Microsoft Japan Co., Ltd.
Tadashi Okazaki conveyed his gratitude for the valuable opportunity to showcase this compelling example of SLM utilization on board aircraft in an offline environment, leveraging the power of Fujitsu Kozuchi for Japan Airlines. He highlighted the project as an innovative and groundbreaking initiative within Microsoft Japan, serving as a clear testament to the high level of technical capabilities and the strong and effective partnership between Fujitsu and Headwaters. Okazaki expressed his firm belief that this project will significantly contribute to the further development of Japan Airlines’ AI initiatives and enhance both flight safety and the renowned Japanese hospitality that JAL is known for.
Future Plans
Building upon the significant success of the field trials, Fujitsu and Headwaters are resolutely committed to continuing their rigorous testing efforts towards achieving full production deployment for JAL. Their ultimate and overarching goal is to seamlessly integrate the generative AI solution into JAL’s existing generative AI platform, creating a unified, cohesive, and highly efficient workflow for cabin crew members.
In addition, Fujitsu is actively planning to expand the inherent capabilities of Fujitsu Kozuchi by strategically incorporating SLMs specifically tailored to a wide range of specific types of work and operational tasks. This expansion will empower organizations across diverse industries to leverage the transformative power of generative AI for a multitude of applications, ranging from providing exceptional customer service to performing sophisticated data analysis.
Both companies remain deeply dedicated to supporting JAL’s ongoing operational transformation through the strategic application of AI, contributing to effective problem-solving, consistently improved customer service, and the proactive addressing of industry challenges. They firmly believe that AI has the potential to revolutionize the aviation industry, making it significantly more efficient, safer, and customer-centric in all aspects.
This powerful collaboration between Fujitsu, Headwaters, and Japan Airlines represents a significant and meaningful step forward in the practical application of generative AI within the aviation industry. By leveraging the power of AI to streamline critical cabin crew workflows, these forward-thinking companies are paving the way for a future where technology empowers employees to provide exceptional service and dramatically enhance the overall passenger experience. The highly successful field trials clearly demonstrate the potential of on-device generative AI solutions to transform operations in a wide range of industries where network access is often limited or unreliable, opening up exciting new possibilities for improved efficiency, increased productivity, and unparalleled innovation. As Fujitsu and Headwaters continue to refine and expand their AI offerings, they are strategically positioned to play a leading and influential role in shaping the future of work across various industries and sectors. The potential for further innovation and improvement in operational efficiency, safety, and customer service remains substantial, promising a future where AI plays an increasingly vital role in the aviation industry and beyond.