FAQ · Dupes and controversies

Can artificial intelligence create a great perfume?

AI can suggest accord candidates, optimize raw material costs, and accelerate brief-to-formula iterations. It does not yet generate a finished niche perfume of recognized artistic stature without substantial perfumer authorship.

The essentials

Artificial intelligence entered fragrance development around 2018, when Symrise and IBM Research announced Philyra, a machine learning system trained on the supplier's internal database of formulas, raw material profiles, and consumer evaluation data. Philyra co-created the first commercially released AI-assisted fragrance, Egeo On You by Brazilian house O Boticario, launched in 2019 (Perfumer & Flavorist, accessed 2026-05-29). Givaudan followed with Carto, a touchscreen tool that lets perfumers compose accords from a curated palette of materials with real-time olfactive guidance.

These systems function as creative assistants rather than as autonomous creators. They propose accord directions, predict cost and stability outcomes, and shorten the iteration loop between brief and submission. The perfumer retains editorial control: choice of palette, narrative intent, balance of materials, and the decision to ship. No major fragrance has been released to date with sole authorship credited to a software system, and the industry has yet to converge on how AI contribution should be disclosed when it occurs.

Whether AI can create a "great" perfume depends on the definition of greatness. For greatness as efficient optimization within a known consumer space, current tools already perform well. For greatness as singular artistic statement, the kind associated with names such as Edmond Roudnitska, Jean-Claude Ellena, or Bertrand Duchaufour, AI remains far from autonomous capability. The systems are trained on what exists, and they optimize toward the known rather than discover genuinely new olfactive territory (Givaudan corporate communications, 2024).

Tools currently used in the industry

The four large suppliers each operate proprietary digital tools. Symrise uses Philyra in collaboration with IBM. Givaudan operates Carto, a touchscreen accord-building system deployed across its perfumery schools. Firmenich, now part of DSM-Firmenich, has discussed its own internal AI tools but has not branded them publicly. IFF has invested in formula data infrastructure that supports machine-assisted brief response.

These tools share three layers: a structured database of materials with olfactive descriptors and performance data, a prediction engine that suggests combinations and forecasts olfactive outcomes, and an interface that allows perfumers to interact with proposals. The pattern is consistent: the system surfaces options the human evaluates and refines.

What AI does well

Three domains show measurable acceleration. The first is reformulation under IFRA constraints. When a restriction lowers the allowed concentration of a material, AI can quickly suggest replacement combinations that preserve the original olfactive direction, a task that previously required weeks of manual iteration. The second is cost optimization. AI can rebalance a formula to remove an expensive material while maintaining olfactive character, valuable for mass-market briefs with tight cost targets.

The third is creative ideation in defined territories. For a brief such as "modern fougere with woody warmth and gourmand accents," AI can propose a dozen candidate directions that a perfumer evaluates and refines. This accelerates the early-stage exploration that often consumes the longest time in traditional development cycles (Fragrantica, accessed 2026-05-29).

What AI cannot yet do

AI cannot originate a brief in the sense of identifying a cultural moment or an unmet olfactive territory and translating it into a fragrance concept. The decision that a fragrance about salt-encrusted rocks at low tide might resonate now is a human cultural reading. AI receives the brief; it does not generate it.

AI also struggles with the singular artistic decision. The choice to push a dose of cumin that most consumer panels would reject because it gives the composition a recognizable signature, the choice of an unusual material that breaks family conventions, the willingness to release a fragrance that is challenging rather than universally liked, these are aesthetic decisions grounded in a worldview. Current AI systems, trained on aggregated consumer data, optimize toward acceptance rather than away from it.

Documented commercial cases

The named cases of AI-assisted fragrances released to date sit predominantly in mass-market territory. Egeo On You by O Boticario (Symrise and IBM, 2019) was the public marker. Coty has discussed AI use in its development pipeline without naming specific launches. L'Oreal Research has published on AI-assisted ingredient palette curation.

Niche perfumery has been slower to declare AI involvement, partly because the value proposition of niche brands rests substantially on the named perfumer as creative author. A house whose identity is built on the singular voice of its perfumer has commercial reasons not to highlight algorithmic assistance, even when it occurs at the supplier level. This silence is itself a data point about how the industry currently positions AI relative to its creative claims (Now Smell This, accessed 2026-05-29).

The niche perfumery question

For niche perfumery, the AI question is sharper than for mass market because the genre's identity claim is grounded in original artistic expression. A niche fragrance attributed to "Alessandro Gualtieri" implies a specific human creative perspective; a niche fragrance attributed to "the algorithm" raises a different commercial proposition.

The likely future is hybrid: AI assists at the supplier level in palette generation, cost optimization, and reformulation, while the niche perfumer retains editorial authorship, narrative intent, and the final accord choices. Whether this evolves toward fully AI-authored niche fragrances depends as much on cultural readiness to value such works as on technological capability.

Sources

  • Perfumer & Flavorist, industry articles on Philyra, Carto and AI-assisted fragrance development, 2018 to 2024. Accessed 2026-05-29.
  • Givaudan corporate communications, Carto and the digital perfumer, supplier documentation, 2024.
  • Fragrantica, editorial coverage of AI-assisted releases including Egeo On You. Accessed 2026-05-29.
  • Now Smell This, editorial commentary on the cultural reception of AI in niche perfumery. Accessed 2026-05-29.
Published 29 May 2026 · Updated 30 May 2026 · Last fact check: 30 May 2026 · Osmetheca · Editorial team