Grupo Boticário and Haut.AI Expand Strategic Partnership to Nearly 4,000 O Boticário Stores

Grupo Boticário expanded Haut.AI’s in-store AI skin-analysis platform from a 24-store pilot to ~4,000 O Boticário locations nationwide in Brazil. Pilot results showed an ~80% increase in average skincare order value, with the system analyzing 150+ facial biomarkers in ~80 seconds to generate personalized recommendations for advisors. The rollout is positioned as one of the largest global in-store AI beauty deployments, with further expansion to other Grupo Boticário brands planned in September.

Analysis

The more important signal here is not the rollout itself, but that in-store beauty is becoming a data moat business: whoever captures consultation data, skin profiles, and repeat purchase history can lift attachment rates and migrate customers into higher-margin skincare. That favors vertically integrated beauty operators with dense store networks and proprietary CRM over pure-play product brands, while pressuring rivals that still rely on static shelf merchandising. The second-order effect is that the competitive bar for premium skincare now includes algorithmic personalization, not just brand equity.

For public markets, the direct earnings impact is likely modest near term, but the playbook is relevant for ULTA, Sephora-owned operators, and prestige skin-care brands such as EL and LRLCY. If AI-assisted consults genuinely raise AOV, the biggest P&L lever is mix: more skincare, fewer discount-driven basket cannibalization effects, and better labor productivity per store visit. The catch is that pilot uplift can be heavily selection-biased; scaling from a handful of enthusiastic advisors to a national field force usually compresses the headline benefit materially.

The consensus risk is overrating immediate monetization and underrating the slower value of data flywheel accumulation. The thesis would be falsified if the next 1-2 quarters show no improvement in ULTA-style attachment metrics, or if beauty retailers cite adoption friction, privacy pushback, or no lift in skincare sell-through. Over 6-18 months, the real winner is whichever retailer turns AI consultation into recurring CRM conversion, not whichever vendor wins the loudest press release.

Key Decisions for Investors

• No direct trade in AIPG/BDRFY/GRO: treat as a private-company validation event, not a liquid catalyst.
• Build a small relative-value long in ULTA on weakness versus XLY/XRT over the next 1-3 months only if upcoming earnings or commentary show skincare mix and conversion improving; otherwise stay flat.
• Use ULTA as a watch item for evidence of store-level AI attachment economics: if management references higher basket size or labor productivity, add on confirmation; if not, fade the narrative.
• Track EL and premium skincare peers for competitive spillover: if they begin marketing similar AI consult tools within 1-2 quarters, it confirms the category is becoming table stakes rather than a moat.
• Set a falsifier on ULTA: if same-store sales do not accelerate or gross margin fails to improve over the next two quarters, assume the AI rollout is a marketing win, not a P&L driver.