Artificial Intelligence

Global perfume houses employ artificial intelligence to innovate scents

BUZZINGA team4 min read

An Agence France-Presse investigation revealed how major global perfume companies such as Estée Lauder and International Flavors & Fragrances are using artificial intelligence technologies to accelerate the development of fragrance formulations, check regulatory compliance, and find alternatives to animal ingredients, without replacing the creative sensibilities of human perfumers.

Machine learning is entering Parisian and global perfume laboratories as a supportive tool that cuts short the examination of thousands of molecular properties and international regulatory constraints, but it remains incapable on its own of measuring emotional response and the interaction of a scent with human skin.

Integrating algorithmic suggestions into laboratories and artistic perfumery practices

An original investigative report published by Agence France-Presse on November 14, 2026, showed how major perfume manufacturers in France, the United States, and Switzerland are employing artificial intelligence technologies in product development stages, ingredient selection, and regulatory compliance verification. The agency conducted field interviews with researchers and perfumers associated with Estée Lauder and International Flavors & Fragrances, alongside smaller perfume houses.

Rather than announcing the replacement of human perfumers with machines, the report meticulously documents how algorithmic proposals are added to long-standing laboratory and artistic practices to accelerate initial empirical research.

Estée Lauder’s innovation center in Paris and resolving molecular diffusion contradictions

The Estée Lauder fragrance innovation center, which opened in the French capital of Paris in 2025, operates across a portfolio of luxury brands that includes Jo Malone London, Le Labo, Kilian Paris, and Tom Ford Beauty. Lucas Nannini, vice president of fragrance innovation at the center, described the use of artificial intelligence models to explore chemically difficult and complex formulas.

Among the most prominent technical challenges facing designers is creating an aromatic oil that has a rich, deep scent while simultaneously being light enough to diffuse into the air. The model is able to search among possible combinations of molecular properties much faster than a human team’s ability to test each option manually in the laboratory.

Regulatory compliance checking across markets and the search for animal ingredient alternatives

Verifying regulatory compliance represents another highly critical practical application, as perfume formulas may involve more than 1,000 potential ingredients whose permitted concentrations vary from one international market to another. Artificial intelligence-backed software helps determine whether a particular substance is permitted in the destination country and the concentration limits applicable to it, which shortens administrative work without replacing official regulatory records, safety tests, and approved system approvals.

AFP also documented the employment of models to search for ingredients capable of replacing animal-derived materials by suggesting molecules with similar properties or close aromatic compounds, while emphasizing that this does not guarantee scent matching, longevity, or human safety profiles without empirical testing and specialized evaluation.

Estimated time savings of 30% and the limits of automation against human creativity

International Flavors & Fragrances reported that its use of artificial intelligence systems trained on years of internal formula data can reduce fragrance development time by roughly 25% to 30%, an industry-issued estimate rather than a total gain proven by independent trials, while competing companies such as Givaudan and Robertet are investing in similar paths.

Conversely, a number of perfumers warned against confusing algorithmic optimization with true artistic creativity. Unexpected variances, emotional responses, and differing scent perceptions from person to person remain elements resistant to digital measurement, and a chemically promising formula may fail to achieve artistic or commercial success.

The entry of machine learning algorithms into major perfume houses represents a practical demonstration of the expansion of artificial intelligence applications into hybrid scientific and industrial sectors that go beyond texts, video, and traditional software. However, the final verdict on the viability of these laboratory investments will depend in coming years on the results of actual products introduced to the markets, the measurement of time and cost savings through independent studies rather than relying solely on manufacturer estimates, as well as rigorous verification of the safety of animal material substitute compounds and the ability of algorithmically proposed formulas to achieve sustainable commercial and artistic acceptance among consumers.

In light of this delicate overlap between molecular chemistry and human artistic sense, perfumers remain steadfast that the machine proposes rapid computational possibilities but does not test the impact of a fragrance on human memory and emotion in real-world environments.

Frequently asked questions

Question: What is the development time reduction percentage announced by International Flavors & Fragrances?
Answer: The company stated that its systems trained on internal formula data can reduce fragrance development time by approximately 25% to 30% according to its own estimates.

Question: How does artificial intelligence help perfume companies with regulatory compliance?
Answer: It helps review formulas that may include more than 1,000 potential ingredients to verify permitted substances and legal concentration limits in each target country prior to formal approval procedures.

Question: Have global perfume houses dispensed with human designers?
Answer: No, interviews confirm that algorithms act as a supportive tool in the laboratory while sensory evaluation, empirical testing, and artistic creativity remain the responsibility of human designers and experts.