By leveraging autonomous AI agents, Shiseido has succeeded in reducing the average time required to search for cosmetic raw materials from 60 minutes to just 3 minutes—a reduction of approximately 95%. This breakthrough is driven by the digital transformation of the entire R&D process and the advanced utilization of research data accumulated over many years.

1.1 The Impact of a 95% Reduction in Search Time and Announcement at the Pitch Contest

At the “Agentic AI Summit ’26 Spring” hosted by Google Cloud on March 19, 2026, Shiseido presented its latest R&D case study featuring autonomous AI agents. Participating in the “5th Generative AI Case Study Award Pitch Contest” during the event, Shiseido’s Mr. Tsurushige and his team unveiled the results of implementing AI agents into the cosmetic raw material discovery process. Previously, researchers spent an average of 60 minutes per search when manually screening materials and searching databases; with the introduction of AI agents, this time was dramatically compressed to just 3 minutes, representing an approximately 95% reduction in task duration. Beyond mere time savings, the technology has vastly increased the probability of discovering diverse material options that human researchers might previously have overlooked. By enabling the rapid examination of a much broader range of raw material data, the speed and quality of idea generation in the lab have evolved significantly. The following image illustrates an advanced research environment powered by AI.

1.2 Challenges in Traditional Material Screening and the Need for AI

In cosmetic formulation development and new product planning, identifying and selecting the right raw materials is a critical step. However, researchers must evaluate a wide array of factors, including chemical properties, safety, skin compatibility, and interactions with other ingredients. Historically, R&D teams relied heavily on individual experience, personal memory, and historical lab notebooks, making the discovery process highly dependent on individual expertise—a long-standing operational bottleneck. Furthermore, given the vast and growing array of new raw materials and potential combinations worldwide, comprehensive manual searches faced physical limits. Even seasoned researchers devoted an average of 60 minutes to a single raw material search, leaving insufficient time for high-value tasks such as conceptualizing new product ideas and driving innovative breakthroughs. The introduction of AI agents offered an indispensable solution, executing rapid, high-precision searches and comparisons to liberate researchers from routine tasks and allow them to focus on creative scientific endeavors.

2. Shiseido’s AI Technologies Powering Formulation Development and Safety Evaluation

2.1 The “VOYAGER” Formulation Development AI and Its Autonomous Search Mechanism

A cornerstone of Shiseido’s AI integration is its continuously evolving formulation development AI, “VOYAGER.” VOYAGER incorporates an AI recommendation feature using proprietary algorithms co-developed with Accenture. In addition to expertise accumulated over more than a century of research, the system has learned from over 500,000 technological and formulation data points, integrating veteran researchers’ tacit knowledge and complex raw material interactions that were previously difficult to digitize. When a researcher inputs desired product concepts or functionality using natural language, the AI autonomously generates tens of thousands of formulation options and narrows them down to a few dozen optimal candidates. Following its full operational deployment in 2024, the summer of 2026 will mark the launch of the first product developed using VOYAGER: a mist-type suncare product offering both scent and color experiences, released under “fibona,” an open-innovation co-creation brand originating from the Shiseido Global Innovation Center. By fusing human inspiration with the AI’s advanced computational capabilities, Shiseido is achieving unique product development that delivers both sensory delight and functional performance.

2.2 Safety and Biodegradability Evaluation AI alongside the Responsible AI Use Policy

Beyond formulation development, Shiseido proactively applies AI technology to evaluate ingredient safety and assess biodegradability to reduce environmental impact. In collaboration with the National Institute of Technology and Evaluation (NITE), Shiseido developed an evaluation system based on the Ministry of Economy, Trade and Industry’s “Degradability AI-QSAR” model under the Chemical Substances Control Law. This system rapidly predicts and extracts toxicological information, such as biodegradability and skin sensitization, directly from chemical structures—drastically reducing the burden of empirical testing that previously demanded extensive time and cost. Consequently, useful raw materials that were previously set aside due to a lack of data can now be utilized. Concurrently, to manage risks associated with cutting-edge technology, Shiseido instituted its “Responsible AI Use Policy” on June 20, 2025. In addition to ensuring regulatory compliance and transparency, this policy emphasizes a “human-centric” approach. By establishing a governance framework where human eyes perform final verifications on the accuracy and ethics of AI-generated data, Shiseido ensures safe and reliable innovation.

3. R&D Transformation and Future Outlook

3.1 Preserving Master Crafts Skill and Unlocking Young Researchers’ Creativity

The most profound transformation brought about by AI agents and formulation development AI is the elimination of key-person dependencies and the “democratization of research.” In cosmetics development, the “tacit knowledge” of veteran researchers—such as the subtle nuances of emulsification and ingredient compatibility—has historically played a vital role despite being difficult to codify in words or numbers. By training AI on both commercial and internal knowledge assets, Shiseido has enabled the permanent preservation of these technical assets. As a result, even less experienced young researchers can now craft high-quality, commercially viable formulations in a short timeframe by building on the AI’s advanced recommendations. The multifaceted approaches and unexpected ingredient pairings calculated by the AI stimulate young researchers’ imaginations, sparking novel ideas that transcend conventional thinking. As depicted in the image below, a new research and development style defined by human-AI co-creation is taking firm root.

3.2 The Future of R&D Innovation in the 2030 Medium-Term Strategy

Under its “2030 Medium-Term Strategy,” Shiseido outlines three key pillars aimed at maximizing brand value and achieving sustainable growth. Within these pillars, expanding investment in AI and digital technologies stands as a top priority for “evolving global operations” and “strengthening value creation capabilities.” Enhancing AI in R&D—highlighted by autonomous AI agents that cut raw material search time by 95% and the VOYAGER platform—directly accelerates development lead times and boosts productivity. Heading toward 2030, Shiseido will not only deepen its presence in core strengths like skincare and suncare, but also expand into broader beauty domains, including Medical & Derma and new market creation. As AI rapidly handles routine searches and calculations, human researchers can dedicate themselves to exploring sensory value, design, and core customer experiences, positioning Shiseido to firmly consolidate its leadership in global cosmetics innovation.

Reference Pages

• [Shiseido Successfully Develops Unique 1st Cosmetic Product Utilizing Formulation Development AI “VOYAGER”] https://corp.shiseido.com/jp/news/detail.html?n=00000000004127
• [Responsible AI Use Policy | Compliance & Risk Management | Sustainability | Shiseido Company, Limited] https://corp.shiseido.com/jp/sustainability/compliance/aipolicy/
• [Shiseido Formulates 2030 Medium-Term Strategy ~Toward a New Growth Trajectory by Maximizing Brand Value~ | PR TIMES] https://prtimes.jp/main/html/rd/p/000002988.000005794.html

【#Shiseido #AIAgent #RawMaterialDiscovery #GenerativeAI #CosmeticsDevelopment #VOYAGER #DX #RandD】