Complete AI Training
Start for Free

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI news ·

Shiseido uses AI to create a water-light mist SPF you actually want to wear

Shiseido's Voyager AI pitched a toner-like mist SPF-light, scented, tinted-showing AI can suggest complete, lab-ready formulas. Faster cycles, fewer dead ends, sensory-first specs.

3 min read
Product development

AI-built, sensory-first mist SPF: what product teams can learn from Shiseido's Voyager

Shiseido used its Voyager AI platform to autonomously propose a mist-type sun care product that feels like a facial toner. It started from a biphasic beauty oil and ended with a light, water-like SPF that blends protection with fragrance and subtle color tone. For product development teams, the signal is clear: AI can move beyond search and actually suggest complete, commercially viable formulations.

Key takeaways

  • Voyager progressed from a Research database (launched in 2024) to an AI system trained on formulations that can propose end-to-end cosmetic recipes.
  • It compiles 500k+ data points, ingredient informatics, emulsification know-how, and researcher expertise into actionable recommendations.
  • The first output: a mist SPF that prioritizes sensory appeal-fresh feel, pleasant scent, and color tone-alongside UV protection.
  • Core consumer insight: people want UV protection that doesn't feel like a chore.

Why it matters for product development

  • Faster cycles: Less manual research means more time shifting between iterations and testing rather than searching.
  • De-risking early stages: Ingredient compatibility and mechanism hints reduce dead-ends before pilot batches.
  • Team leverage: Junior formulators can produce commercially ready concepts with AI assistance and expert review.
  • Sensory as a spec: Fragrance, color tone, and feel move up front-treated as requirements, not late-stage add-ons.

Behind the screens: how Voyager works

Voyager absorbs cosmetic-specific knowledge including emulsification mechanisms, ingredient interactions, and lessons from past research. It turns that into formulation proposals aligned with a consumer profile-here, a toner-like mist experience with SPF.

Ingredient informatics helps the system reason about individual components and their interactions. The outcome is a shortlist of viable bases, process guidance, and constraints a lab can execute and refine.

Sensory-first SPF: what's new

  • Format: Mist that applies like a toner, not a thick lotion.
  • Base: Originated from a biphasic beauty oil concept, translated into a lighter, water-like feel.
  • Design integration: Fragrance and color tone built alongside UV performance rather than tacked on later.
  • Emotion-led brief: "Protection without the chore" as the guiding insight.

How to adapt your development process

  • Turn sensory into measurable inputs: Define targets for spreadability, volatility profile, droplet size, residue score, scent intensity, and transfer/staining risk.
  • Co-create with AI: Use an AI system to generate multiple formulation paths, then pressure-test in the lab for stability, sprayability, and SPF readouts.
  • Front-load compatibility: Screen UV filters against solvents, film formers, fragrances, and pigments early to avoid phase separation and nozzle clogging.
  • Process-aware specs: Include shear ranges, order of addition, and cooling profiles in the proposal; they matter as much as the ingredient list.
  • Consumer loop: Validate "non-chore" metrics: dry-down time, tackiness, scent persistence, and reapplication willingness.

Technical checkpoints for a mist SPF

  • Uniform filter distribution: Ensure filters remain evenly dispersed or dissolved to hit label SPF across sprays.
  • Spray system fit: Match viscosity and surface tension to pump spec; check aerosol vs. non-aerosol options and regional rules.
  • Stability: Thermal/light stability, phase behavior of biphasic systems, fragrance/color stability, and packaging interactions.
  • Aesthetics: Color tone should not stain skin or fabric; fragrance should not overpower or conflict with actives.
  • Testing: SPF/UVA testing per region, water resistance where relevant, and repeatability across lots.

Data you'll want in your own AI stack

  • Ingredient-level properties: polarity, volatility, refractive index, odor thresholds, and interaction flags.
  • Process metadata: shear, temperature, time, and order-of-addition tied to outcomes.
  • Sensory panels: structured ratings linked to exact compositions and process notes.
  • Regulatory tags by market to keep proposals compliant from day one.

Metrics to track

  • Concept-to-pilot time and number of iterations per launch.
  • First-pass stability and first-pass SPF hit rate.
  • Consumer reapplication intent and daily-use adoption.
  • Cost-to-quality ratio vs. prior baselines.

Launch timing

The new mist sun care product is slated for summer 2026 under fibona, Shiseido R&D's open innovation program.

Level up your team's AI fluency

If you're building similar AI-assisted workflows for R&D and product, structured training helps shorten the learning curve. Explore practical options by role here: AI courses by job.

Share