How Dermalogica and HelloBiome Are Using AI to Personalize Beauty

When HelloBiome founder Dr. Elsa Jungman learned of the microbiome — or the trillions of microorganisms that live on the body and contribute to one’s overall health — she immediately sought to leverage data on the discovery, only to realize very little had been conducted.

“There was no diverse panel of data to understand how our [microbiome] compositions differ,” Jungman said in a conversation about the impact of AI on beauty with Dermalogica global chief executive officer Aurelian Lis moderated by Oliver Chen, managing director and senior equity research analyst at TD Cowen.

So Jungman sought to create the database she needed, harnessing AI to develop at-home microbiome testing kits — the findings of which have been used by brands including K18 and Credo Beauty to inform product development.

“We use AI to train what we call a ‘classifier,’ which tries to understand skin segmentation beyond the traditional dry, oil skin types; we’re trying to understand microbiome sequencing information,” said Jungman, whose next step is using AI-powered predictive models to “develop the best solution for each of those [skin] patterns, and also predict which new ingredient or formula will have the best efficacy.”

While Dermalogica is similarly using AI to inform formulation processes, it is also tapping the technology to streamline internal operations.

“What AI has done is help democratize technology across the organization,” Lis said. “We believe, very optimistically, that this is an assistant for everybody — every department — in the company.”

The technology isn’t, he said, a replacement for analytics platforms like Tableau or resource planning softwares, but rather — and perhaps contrary to the beliefs of some — about bringing more emotion and personalization into brand development.

“AI is going to up the standard, because every company is going to be able to personalize, the fact that [a consumer’s] name is on the top righthand corner of a website is going to mean less and less — the differentiating factor of what pure luxury is going to be is going to come back to human touch,” Lis said. “How to make that human touch exciting with AI is where we need to be spending our time.”

Importantly, as Chen put it, “training is the secret sauce of AI” — and those wielding the technology bear the ethical responsibility of training it without imparting bias, conscious or unconscious.

“There are new tools, shopping assistants cropping up everyday and sometimes I will ask founders, ‘How did you train that model?,’ and they can barely tell me,” Jungman said.

“The company approach that makes me quite fearful is that, let’s just get all of our data into a data lake and not define what we’re going to use it for,” Lis said. “When you want to do something, you start with a question — you don’t start with the answer.”

Dermalogica is taking a use case-specific approach to training its AI models, uploading trend reports, email exchanges and even a year’s worth of customer service calls which “we use to detect not just what [the consumer] is talking about, but also the emotion behind the call,” Lis said.

“The biggest risk is that we as an industry don’t adopt AI and lean into the technology,” he said.