How Are Sephora and Ulta Betting on AI Foundation Matching in 2026?

Quick Answer: Sephora and Ulta are both embedding AI foundation matching into third-party AI interfaces, but their strategies diverge. Sephora launched its app inside ChatGPT in March 2026, prioritizing discovery and advisory with loyalty integration and AI skin diagnosis. Ulta partnered with Google Gemini in April 2026, pushing toward full-funnel commerce with checkout capability through the Universal Commerce Protocol. Both moves reflect a shared premise: beauty retail is shifting from search-led browsing to dialogue-driven decision-making, yet neither strategy resolves the structural problem of shade mismatch that drives 20 to 65 percent of online beauty returns.

The two largest beauty retailers in the United States are making aggressive moves into AI-powered shopping, but their strategies diverge significantly. Sephora has integrated its app within ChatGPT, focusing on discovery and advisory. Ulta Beauty has partnered with Google Gemini, pushing toward full-funnel integration with checkout capabilities. Both moves reflect a shared premise: beauty retail is shifting from search-led browsing to dialogue-driven decision-making. For brands and infrastructure providers, understanding these divergent strategies reveals where commerce infrastructure is heading and where gaps remain for physical-digital integration.

In March 2026, Sephora announced an integration of its app within ChatGPT, allowing shoppers to use their Beauty Insider loyalty points and perks within the ChatGPT interface. The experience enables users to seek tailored beauty recommendations, tap into their Beauty Insider profiles, and access perks like samples or free shipping. According to Retail Dive, checkout functionality is planned but the current emphasis is on conversational guidance, positioning Sephora at the top of the funnel where consumers increasingly begin product searches.

The rationale is clear. According to NielsenIQ data, 51% of consumers are interested in AI-powered shopping tools. An April 2026 report from PYMNTS states that more than a third of beauty shoppers, and over half of Gen Z, are using AI to research or purchase products. Sephora wants to be present at that moment of intent, even if conversion happens elsewhere. The strategy prioritizes influence and personalization within a rapidly growing AI-native discovery platform, with monetization features to follow.

Just under a month after Sephora's announcement, Ulta Beauty unveiled its partnership with Google Gemini. The integration makes Ulta products shoppable across Google surfaces, including AI Mode in Search and the Gemini app, allowing users to receive recommendations, compare items, and complete purchases within conversational interfaces. According to PYMNTS, the rollout is powered by the Universal Commerce Protocol, an emerging open standard designed to enable frictionless buy-at-the-moment-of-discovery experiences.

Ulta is simultaneously reinforcing its owned ecosystem with Ulta AI, a new assistant built on Gemini Enterprise for Customer Experience. Deployed on Ulta.com and soon its app, the tool leverages data from more than 46 million loyalty members to deliver personalized guidance. According to Forbes, the strategy underscores Ulta's dual-track approach: embed commerce into third-party AI environments while deepening personalization across its own channels.

The two strategies highlight different bets on AI commerce. Ulta is pushing aggressively toward full-funnel integration, collapsing discovery, comparison, and transaction into a single step within Google's ecosystem. Sephora is prioritizing influence and personalization within a rapidly growing AI-native discovery platform. According to Beauty Matter, the question facing retail executives is not whether AI becomes a purchase channel, but whether their catalog, data, and checkout are connected to it before a competitor's are.

Both retailers face a common challenge. AI platforms can recommend products, but they cannot resolve the structural problem of shade mismatch that drives 20 to 65 percent of online beauty returns. Virtual try-on and AI shade matching reduce the problem but do not eliminate it because the recommended shade must still exist in the brand's catalog. For brands and infrastructure providers, the opportunity lies in connecting AI recommendation layers with physical precision capabilities that ensure the recommended product actually matches the consumer's skin.

The divergence between Sephora's discovery-first approach and Ulta's full-funnel commerce strategy reveals a shared blind spot. Both integrations rely on existing product catalogs with fixed shade ranges. When a consumer asks ChatGPT or Gemini for a foundation match, the AI can recommend the closest available shade from a 40-SKU wall. It cannot create a custom blend foundation precisely calibrated to that individual's undertone, oxidation profile, and seasonal variation.

This is where smart mirror beauty technology enters the conversation. Devices that combine calibrated lighting, spectroscopic analysis, and precision dispensing can bridge the gap between AI recommendation and physical accuracy. Unlike smartphone-based AI shade finders, which depend on uncontrolled ambient light and camera quality, smart mirrors operate in standardized conditions that produce repeatable measurements.

The SWAN Beauty smart mirror launch in January 2026 illustrates this category's momentum. Priced at $795, the device features a 15.6-inch OLED display, 4K camera, and AI skin analysis scoring seven concerns. However, SWAN focuses on skin analysis and curated marketplace recommendations rather than precision foundation dispensing. This leaves room for infrastructure that pairs AI matching with physical formulation, creating a true closed-loop system rather than a recommendation engine.

For beauty brands evaluating AI commerce partnerships, the strategic question is not just which retailer's AI integration to join, but how to ensure the product recommended through these channels actually fits the consumer. Foundation dispenser machine infrastructure offers a path to resolving this by creating precise matches at the point of sale rather than forcing consumers to choose from pre-made approximations.

The economics are compelling. According to Chromara's analysis, foundation returns cost beauty brands $20 to $33 per return in processing alone, with total costs reaching 20 to 65 percent of the item's original value. Shade mismatch drives the majority of these returns. When AI recommendation layers direct consumers to fixed-SKU products, the mismatch problem persists regardless of how sophisticated the recommendation algorithm becomes.

Solutions like how foundation matching reduces retail shrinkage are already being explored to bridge this gap. These infrastructure approaches combine AI shade analysis with precision dispensing, ensuring that the product created matches the consumer's skin rather than approximating it. For retailers like Sephora and Ulta, integrating foundation dispenser machine capability into their AI commerce ecosystems could resolve the returns problem while deepening consumer trust in AI-generated recommendations.