Brazilian cosmetics group O Boticário is using artificial intelligence technology for fragrance development in a project that aims to optimize specialist human work and improve customer engagement.
O Boticário is the second biggest Brazilian cosmetic firm – the first is Natura, the owner of The Body Shop and Aesop brands. It is the largest cosmetic franchise in the world with revenues of $3.7 billion.
In a world first for the sector, the company is using Phylira, a platform developed by IBM Research in partnership with German producer of flavours and fragrances Symrise and Boticário. The software uses advanced automatic learning algorithms to identify new patterns and new combinations of fragrances that provide closer matches with what customers want.
Fed with millions of information items from perfume formulas and ingredients to perfumery history and consumer acceptance rates, the platform cross-referenced the data that resulted in two new products, to be launched on a pilot basis in early 2019.
“The fragrances are unusual and combines scents of fruits, flowers, sweets, spices, types of wood and even cucumber and condensed milk – ingredients that sometimes go unnoticed by perfumers since they are not very common,” says Tiago Martinello, research and development head at O Boticário.
According to Martinello, use of AI allowed a reduction of development time of the new products from three years usually spent on conventional processes down down to six months: “This shows us how innovation is important in terms of productivity.”
Other AI applications may come on stream as a result of the initial endeavor as part of Boticário’s “operations 4.0″ program, with potential areas of application including quality control, industrial production and supply chain management, as well as other areas outside operations such as sales and marketing.
“There are several other technological resources [in addition to AI] that can be applied in an industry to reduce repetitive jobs and operational tasks and we are studying these possibilities,” Martinello says.
Organizational challenges
Given that AI is still a novelty of sorts in most companies, the organizational challenge can be significant – this was also the case at O Boticário around the fragrance development project, according to Martinello.
“When developing an AI system, a lot of knowledge is required from several different areas and the integration of these teams is a big challenge,” the executive says.
When it comes to innovation, engaging key people from the board and to direct participants of the project is another challenging aspect of the job, according to Martinello, as well as improving awareness of the topic.
“When it comes to using AI, the need for engagement was very evident because it is a new topic and sometimes creates controversial understandings.
At a technical level, Martinello highlights that the job of “teaching” the system was very extensive: “it takes a lot of training and improvement until it reaches a level of delivery that is compatible with the company’s quality standards.”
Enhancing soft skills
The first part of the process of AI-aided fragrance development is related to technology, says Martinello. At this stage, formulas, raw materials, success stories and trends are studied by Phylira, so that its algorithm detects patterns and provides for unprecedented combinations.
Humans take over in the following stage, where product development professionals analyze the information and product prototypes. Martinello argues that the optimization of the first stage allows more time for the creative part of the process.
“Artificial intelligence is yet another tool available to perfumers to optimize fragrance development, but products are born out of briefings thought out by experts on consumer behaviors and demands, as well as marketing and R&D and perfumers themselves – they all add to that mix before the product reaches consumers,” he adds.
While the company sees use of AI as a way to enhance the work done by its fragrance specialists, it still cannot predict a near future where soft skills would be replaced by AI in its production processes.
“What [AI ] already provides is support to [soft skills] – the development of fragrances aided by this technology the evidence of this. It’s a complement rather than a substitute.”
