At CEATEC 2025, Fujitsu expanded its ongoing exploration of what it officially calls its “AI technologies for human augmentation” with a new Golf Motion Analysis Experience, powered by the Kozuchi AI platform and developed in collaboration with Uvance partner AIGIA’s golf swing–analysis app. This builds on last year’s showcase, where the company demonstrated how Kozuchi’s advanced skeleton recognition could analyze basketball shooting mechanics, yoga postures, and lifting movements in industrial settings. The 2025 exhibit continues that progression, showing how the same AI engine can be quickly adapted to interpret a completely different sport —one that requires precise, highly nuanced biomechanics.
During the show, a Fujitsu engineer provided insight into the technology behind this year’s demo. He explained that the system is not intended to function as a complete golf simulator, but rather to demonstrate how Kozuchi can understand real-world biomechanical movements and provide performance-oriented feedback through an AI agent. He described how last year’s basketball demo used the same software foundation, emphasizing that golf reinforces the platform’s ability to work across different sports and physical activities.
A Natural Evolution From Basketball and Yoga to Golf
The 2025 golf demo focuses on two motions that are notoriously challenging even for experienced players: bunker shots and approach shots. Both require controlled swings, precise weight transfer, and a stable body position—ideal characteristics for a system designed to analyze biomechanics.
According to Fujitsu, the significance of the demo is that no new motion-analysis engine had to be built. The golf experience runs on the same Kozuchi core used for basketball and yoga. Since the platform is designed to detect subtle variations in posture, rotation, timing, and balance, adapting it to golf required only the integration of expert golf-specific reference data.
Kozuchi is designed to be sport-agnostic, capable of interpreting a wide range of activities as long as they involve measurable body movement. For Fujitsu, this versatility supports its strategy to position these AI-driven applications as human-augmentation technologies rather than standalone sports tools.
Inside the CEATEC Golf Booth: A Real-World Challenge, Reconstructed
To make the experience meaningful, Fujitsu recreated the infamous bunker on the 18th hole of the Tokyo Seven Hundred Club’s West Course, known for its steep 1.6-meter rise and notoriously difficult escape angle. Reproduced at life-size scale, the bunker turned the booth floor into a miniature version of one of Japan’s most challenging greenside scenarios.
Visitors attempted a bunker shot while the system captured and analyzed their motion. The experience unfolded in two stages: a practice shot, followed by a final shot where the AI compared both attempts and visualized changes.
Uvance partner AIGIA Integration
Uvance AIGIA refers to the collaboration between Fujitsu Uvance and AIGIA Inc., where AIGIA is a Uvance partner.
The 2025 golf experience incorporates technology from Uvance partner AIGIA, integrating AIGIA’s specialized golf-swing analysis app with Kozuchi’s advanced skeleton recognition AI. This “powered by Uvance” collaboration brings together Fujitsu’s broad human-motion analytics (HMA) with AIGIA’s domain-specific golf expertise.
In practice, the system integrates Kozuchi’s motion capture and biomechanical interpretation with AIGIA’s swing-pattern insights, together with data on the bunker and green conditions, to provide personalized, context-aware guidance through an AI Caddie agent.
The result felt more like a sophisticated training system than a booth attraction. Fujitsu also published a 3D reconstruction of the bunker setup on Sketchfab, showing how sensors, spatial design, and visualization systems work together.
How The Technology Breaks Down a Golf Swing
In the demo, the software tracks posture, joint alignment, upper-body rotation, and weight distribution, then compares these factors with an ideal movement model for the shot. According to the engineer, the entire process operates as a single continuous workflow that includes full-body skeletal tracking, biomechanical interpretation, expert-model comparison, and avatar-based visualization, all processed in real time.
Kozuchi’s primary strength lies in its ability to interpret movement rather than simply capture it. The platform pinpoints how a swing deviates from the optimal version and identifies the underlying cause of those deviations. This interpretative capability is central to its usefulness in training, rehabilitation, and workplace applications.
Beyond Golf: A Platform With Multi-Industry Applications
Although presented through golf, Fujitsu stresses that the underlying platform is not limited to sports. The same engine is used for digital health, post-injury rehabilitation, factory ergonomics, and professional sports judging, such as gymnastics.
Fujitsu works with medical partners to identify movement patterns linked to chronic pain, instability, or physical asymmetry. In industrial environments, the technology is already used to analyze repetitive lifting motions and evaluate long-term strain risk. In both cases, the platform uses the same skeletal-tracking and motion-analysis backbone as the golf demo.
Fujitsu aims to build a single adaptable AI framework capable of interpreting movement across sports, medicine, and workplace environments. The golf demo is not an isolated project but a demonstration of how easily the platform can transition between use cases.
Fujitsu showcased a few applications of its Human Motion Analytics technology at CEATEC 2024.
Kozuchi’s Role in Fujitsu’s AI Roadmap
Kozuchi is Fujitsu’s cloud platform for secure, enterprise-grade AI services. Although motion-analysis demos are the most visually engaging examples showcased at CEATEC, the platform also includes digital twins, simulation pipelines, AI agents, and predictive analytics. Fujitsu uses CEATEC to highlight the human-centered side of its AI technologies, because motion analysis offers a concrete demonstration of how the system interprets real-world behavior.
While Fujitsu did not make any formal comparison between the 2024 and 2025 demos, this year’s presentation suggests a shift toward consolidating different use cases under a single motion-analysis model. The golf experience highlights how Kozuchi can adapt to varied physical activities using the same core engine, illustrating potential pathways for applications in sports training, digital health, rehabilitation, and workplace safety.
Filed in . Read more about Agentic Ai, AI (Artificial Intelligence), CEATEC, CEATEC 2025, Computer Vision, Fujitsu, Golf and Japan.