Haut.AI cuts skin efficacy study setup from months to days with AI platform

Haut.AI has unveiled its Clinical Studies Software, a platform that scores skin efficacy studies’ results with AI rather than human panels. The AI-powered skin and hair analysis company says participant cohorts can run from 100 to 7,000 individuals. Moreover, studies will only take two three days from setup to start.

The platform targets a method that has seen little change in decades. Haut.AI points out that a traditional skin efficacy study typically enrolls 30 to 35 participants and takes eight to 16 weeks to set up. 

“Clinical research in beauty and skin care has reached an inflection point,” says Anastasia Georgievskaya, CEO and Co-Founder at Haut.AI. “The industry has incredible expertise in clinical science, but the tools used to collect and analyze data have remained largely unchanged for years.”

Solving data clutter 

According to the company, the current system requires trained experts to score each image in a skin efficacy study manually. Haut.AI adds that with this step, the same image can draw different scores from different experts, or even from the same expert on a different day.

Haut.AI says that variation enters the dataset as noise — or informational clutter — and notes that noise can obscure the findings on a product’s effects. 

However, it says that an AI model applies one standard to every image universally across sites and timepoints. Haut.AI underscores that its platform’s consistency lets a study span multiple markets without the need to send trained graders to each location.

Standardizing skin scanning

The company commissioned an evaluation of its measurement technology at the Institut d'Expertise Clinique in Lyon, France, comparing it against a consensus panel of expert graders. Haut.AI says the models returned intraclass correlation values of 0.97 to 0.98 for repeatability across five facial endpoints, a level the company describes as excellent by clinical standards. The measure indicates notable consistency in the AI’s repeated readings.

Furthermore, the company says those values hold across all three capture settings that the platform supports, including a clinical camera, a smartphone at a research site, and an at-home selfie. This remained true across participants aged 20 to 70 years old with balanced gender representation.Haut.AI’s CEO says AI grading applies one standard across every image, site, and timepoint.

On the question of how AI compares with the human panel, Haut.AI says its grades correlate strongly with the dermatology consensus for structural aging signs and pigmentation. It does not publish a figure for that comparison.

Haut.AI says the platform quantifies 48 validated biomarkers spanning pigmentation, wrinkles, texture, acne, redness, and pore appearance. Participants need a link rather than an app, and capture runs through the company’s Live Image Quality Assurance technology, which checks position, lighting, and framing before the image is taken.

“Visible skin aging has always been harder to quantify with the same rigor we apply to molecular aging markers,” says Dwaraka. “Haut.AI’s Clinical Studies Software gave us standardized, image-derived measures of facial aging traits that we could pair directly with our DNA methylation data.” 

“That combination let us treat visible aging as a quantitative trait alongside our biological measurements, instead of relying on subjective grading.”

Haut.AI argues that scoring images to a single standard makes visible skin traits usable as quantitative variables alongside molecular data.

What stays human 

Additionally, the company says it is not positioning the software as a replacement for a clinical research organization (CRO). Ethics approval, Institutional Review Board submission, protocol adherence, and participant recruitment remain with the research partner. However, the company says it can train an existing internal CRO team on the platform to operate the software.

“R&D teams that can measure continuously across larger populations don’t just do better science; they make faster decisions,” says Georgievskaya.

Haut.AI says that facial imagery used in analysis is anonymized through patented Skin Atlas technology.

“Our goal is not to replace clinical studies,” concludes Georgievskaya. “It’s to enhance them by making skin measurable at scale.”