Inside the AI Beauty Counter: How Ulta and Startups Are Building Digital Beauty Advisors

The beauty aisle is becoming a software experience. Between Ulta’s push into Ulta AI, rapid improvements in virtual try-on, and startups using machine learning to read skin, shade, and shopping intent, the future of beauty retail looks less like a shelf and more like a digital consultant. For shoppers, that means faster discovery, fewer bad buys, and more personalized recommendations based on what you actually need—not just what is trending on TikTok. It also means retailers are learning to use first-party data to create experiences that feel helpful rather than pushy.

What makes this shift so important is that beauty is one of the most personal categories in retail. Shade matching, skin sensitivity, hair texture, fragrance preference, and budget all shape the final decision. That is why the most effective AI beauty advisors are not just chatbots; they are systems that combine loyalty data, product knowledge, image analysis, and shopping behavior into something closer to an agentic style assistant. As Ulta’s leadership has suggested, the opportunity is not only to improve conversion, but to bring in shoppers who may never have walked into the store in the first place.

For readers trying to shop smarter, this guide explains how the technology works, what Ulta is doing differently, how startups are building competing tools, and what it all means for everyday beauty routines. If you are interested in the broader retail logic behind these changes, it is useful to compare it with other value-driven shopping strategies like value bundles and even the quality checks behind too-good-to-be-true bargains, because AI-driven beauty shopping is ultimately about reducing risk before purchase.

1) Why Beauty Retail Is Ripe for AI

Beauty decisions are high-stakes and highly personal

Buying mascara is not the same as buying detergent. A beauty product can fail because of undertone mismatch, fragrance sensitivity, ingredient intolerance, or simply the wrong texture for your skin type. That complexity creates a perfect use case for recommendation systems, because shoppers often need guidance before they are willing to spend. AI can compress what used to be a long cycle of trial, error, and returns into a more confident first purchase.

There is also a trust problem. Many shoppers have been burned by glossy marketing claims that do not match reality. This is why retailers investing in beauty tech are increasingly focused on data-backed suggestions, real-time feedback, and more transparent product education. When shoppers feel informed, they are more likely to buy and less likely to abandon the basket.

Consumers increasingly start with AI, not search engines

Ulta executives noted at NRF that a growing share of consumers now begin shopping journeys using AI tools like ChatGPT. That is a major behavioral shift. Instead of typing “best foundation for oily skin,” shoppers are asking more nuanced questions like, “What is the best medium-coverage foundation for acne-prone skin in humid weather under $40?” AI assistants can interpret those layered needs much better than classic search.

For retailers, this means discovery is moving upstream. Brands that understand how to appear in AI-led shopping journeys will have an edge, especially as consumers increasingly expect instant, personalized shortlists. If you want to think about how shopping behavior changes when a digital tool interprets intent, the same logic appears in other categories too, such as budget-friendly essentials or tools that actually save time.

Retailers need a system, not a gimmick

The winners will not be the companies with the flashiest demo. They will be the retailers that connect AI to inventory, loyalty data, sampling, and customer service. In beauty, the recommendation is only valuable if the product is in stock, the shade is real, and the customer can trust the advice. That is why first-party data matters so much: it lets brands ground the AI in actual consumer behavior rather than generic internet trends.

That also makes governance essential. Just as organizations build oversight for other sensitive AI systems, beauty retailers need rules for what data is used, how recommendations are generated, and when a human should step in. A useful analogy comes from building an AI governance layer: the smartest systems are the ones that can be trusted at scale.

2) Ulta’s AI Strategy: Loyalty Data Meets Digital Consultation

46.7 million loyalty members are the real engine

Ulta’s biggest advantage is not just store count; it is data. With 46.7 million loyalty members, the company has a large, behavior-rich dataset that can inform product recommendations, category affinities, replenishment timing, and personalization rules. That first-party data can reveal everything from the brands a customer browses to the categories she buys repeatedly and the time of year she updates her routine.

This is important because beauty shopping is seasonal and cyclical. A customer might need a richer moisturizer in winter, a lighter base in summer, and a fragrance refresh before the holidays. AI can spot those patterns faster than a human associate can, especially when the customer shops online and in-store interchangeably. The result is a more useful digital advisor that feels like it remembers your beauty routine.

Agentic AI could become a digital beauty consultant

Ulta leadership has said it sees potential in agentic AI—systems that do more than answer questions and can actively help complete tasks. In a beauty context, that could mean an assistant that builds a regimen, checks compatibility across products, flags ingredients to avoid, suggests alternatives if a preferred shade is out of stock, and maybe even schedules salon services. That is much closer to a digital consultant than a static recommendation engine.

The real commercial upside is not just conversion; it is basket quality. If an AI assistant helps a customer buy a cleanser, serum, SPF, and setting spray that work together, the retailer creates a better experience and a stronger relationship. That is why the concept matters for both revenue and retention. It mirrors what strong post-purchase programs do in other industries, such as client care after the sale, where ongoing support is often more valuable than the first transaction.

International expansion and innovation can reinforce each other

Ulta’s growth plans are not limited to the U.S. The company has spoken about expansion in the UK, Mexico, and the Middle East while continuing to scale its store footprint. That matters because international markets often push companies to standardize processes while still localizing recommendations. AI can help solve that tension by adapting product assortments, languages, climates, and consumer preferences at scale.

In other words, the same intelligence layer that powers a better app experience can also support market expansion. For shoppers, that can mean more relevant assortments and better product education. For the company, it means an infrastructure that can support both growth and personalization without endlessly adding headcount.

Pro Tip: The best AI beauty advisor is not the one that knows the most products—it is the one that knows which products fit your skin, budget, and routine, and then explains the trade-offs clearly.

3) How Virtual Try-On Is Changing the Confidence Equation

AR try-ons reduce one of beauty’s biggest pain points

Virtual try-on tools solve a major friction point: uncertainty. In categories like lipstick, foundation, brow products, and hair color, many shoppers hesitate because the in-cart image is not enough to predict real-life results. AR and camera-based tools let shoppers preview shades on their own face or compare finishes before buying, which can dramatically improve confidence.

That confidence matters most for first-time buyers. If a shopper has never used a particular brand, a try-on experience offers a kind of digital sampling. It will not replace physical testing in every case, but it can narrow choices from 20 shades to 3, making the final decision much easier. In practice, this can lower return rates and shorten the time between discovery and purchase.

What the technology can and cannot do

Virtual try-on works best when the input is good. Lighting, camera quality, and skin tone calibration all affect the accuracy of the experience. A tool may be excellent at suggesting a lipstick family but weaker at predicting how a foundation oxidizes after several hours. That is why AR should be treated as an assistive layer, not a magic mirror.

Retailers that communicate those limits honestly tend to build more trust. Shoppers do not need perfection; they need a tool that improves odds. The same logic applies to other high-consideration purchases like vision insurance or flight pricing, where clarity and expectation-setting reduce frustration.

Where the most useful try-ons live

The strongest try-on experiences usually sit at the point of decision: product pages, shade comparison tools, and mobile app journeys. They also work well when paired with ingredient education and reviews. In beauty, a good AR tool should not just show the product; it should explain finish, coverage, wear time, and skin-type compatibility. That is how a novelty feature becomes a conversion tool.

For shoppers who want the smartest possible beauty purchase, try-on should be combined with practical shopping habits: test in daylight, compare undertones, and check whether the retailer offers easy exchanges. It is a lot like learning to spot a real bargain in a fashion sale—you need both the visual and the practical check before you commit. For a helpful parallel, see our guide on spotting a real bargain.

4) What Startups Are Doing Differently from Big Retail

Startups are often more specialized

While major retailers are building broad ecosystems, startups often zoom in on one problem: skin analysis, shade matching, ingredient matching, or conversational product discovery. That specialization allows them to move quickly and create highly tailored tools. Some use face mapping to estimate tone and texture; others use questionnaires paired with purchase history to offer personalized regimens.

This specialization is valuable because beauty is a trust category. A startup that can solve a single painful issue really well—like finding the right concealer undertone or building a sensitive-skin routine—can earn loyalty fast. In many cases, the best digital consultant is the one that is narrowly excellent, not broadly mediocre. That is similar to how smart consumers evaluate niche services in categories like choosing the right repair pro or finding value bundles.

Conversational shopping is replacing static filters

One of the most promising startup moves is conversational commerce. Instead of forcing users to apply filter after filter, AI can ask follow-up questions the way an in-store consultant would: What is your skin concern? Do you prefer dewy or matte? Are you trying to reduce steps? What is your budget? That creates a more human-feeling experience and helps shoppers express needs they might not know how to sort into dropdown menus.

For everyday shoppers, this means less scrolling and fewer mismatched purchases. It also supports more inclusive beauty shopping because the system can ask better questions about skin tone, hair texture, fragrance sensitivity, and routine complexity. A well-designed chatbot becomes less like search and more like guided consultation.

Startups can also move faster on testing

Because startups are smaller, they can experiment more aggressively with UX, prompts, and recommendation models. They can A/B test whether shoppers prefer “best for you” suggestions or ingredient-led explanations, whether video overlays outperform static shade charts, and whether a regimen builder beats a simple ranking list. That speed often lets them identify high-performing workflows before the big players roll them out broadly.

Of course, speed only matters if the output is accurate. The most successful teams usually pair AI with strong data discipline. Beauty retailers can learn from operationally minded businesses that use data to improve workflows, such as AI-driven order management or structured decision systems like AI governance.

5) First-Party Data Is the Competitive Moat

Why first-party data beats generic training data

In beauty, generic internet data can only take you so far. First-party data tells a retailer what its own customers actually buy, browse, return, repurchase, and review. That means recommendations can be based on real outcomes rather than broad assumptions. A retailer that knows a customer buys fragrance minis every month and switches base products seasonally can create far better suggestions than one that only knows her demographic profile.

That is why first-party data is increasingly a strategic asset. It supports personalization, inventory planning, and retention at the same time. It also helps retail media and on-site search become more relevant, because the system can prioritize what a customer is likely to need next rather than what is merely popular overall.

Data quality matters more than data volume

More data is not automatically better. If data is fragmented across apps, stores, receipts, salons, and returns, the model may struggle to create a coherent customer view. Retailers need clean identity resolution, clear consent flows, and strong data hygiene to avoid garbage-in, garbage-out recommendations. Without that, AI may simply reinforce bad assumptions or push products that do not fit the shopper’s real use case.

This is where transparency becomes a trust signal. Brands should explain why a product is recommended: because of skin type, concern, finish preference, prior purchase behavior, or feedback from similar users. That explanation helps shoppers understand the logic and makes them more willing to try the advice. If you want to see how transparency shapes value perception in another category, the logic is similar to understanding what you pay for in craftsmanship-driven purchases.

Consent and privacy are not optional

Because beauty data can include face scans, skin images, and purchase history, consent must be handled carefully. Retailers should disclose what is collected, how it is used, and how long it is stored. Shoppers should be able to opt out of nonessential data use without losing the basic shopping experience. That is especially important as AI becomes more embedded in recommendation flows.

Think of consent as part of the product, not a legal afterthought. Good privacy design can increase trust and reduce abandonment, while sloppy design can quickly turn a “helpful advisor” into a creepy one. For teams building this kind of system, the same discipline used in sensitive workflows such as airtight consent workflows is worth studying.

6) What This Means for Everyday Shoppers

You should expect better matches, not just more ads

The best consumer outcome of beauty AI is not more upselling. It is better matching. A strong AI advisor should help you avoid buying a serum that clashes with your routine, a foundation that oxidizes badly, or a fragrance that sounds lovely on paper but gives you headaches. In short, it should save you time and money by reducing failed experiments.

That means shoppers can start to ask better questions. Instead of “What is trending?” ask “What solves my concern?” Instead of “What is popular?” ask “What is appropriate for my skin type, budget, and climate?” If an AI tool cannot answer those questions clearly, it is probably not yet useful enough.

Look for systems that explain the why

The most trustworthy AI beauty advisors show their reasoning. They should tell you why a cleanser was recommended, why a shade range fits your undertone, or why a product was excluded. Explanations matter because beauty is full of hidden variables like climate, hormonal changes, and ingredient sensitivities. A transparent system helps you make informed trade-offs.

Shoppers should also compare AI suggestions with reviews, ingredient lists, and return policies. AI is helpful, but it should not replace basic due diligence. That is true in beauty just as it is in travel, shopping, or home projects where smart buying still depends on solid judgment and comparison.

Use AI as a shortcut, not a substitute for awareness

AI can speed up routine decisions, but you still know your own skin best. If you have eczema, rosacea, active acne, or fragrance sensitivity, treat AI suggestions as a starting point and check ingredient labels carefully. The same goes for shade matching: a camera preview can be helpful, but natural light and real-world wear remain important.

For shoppers managing busy lives, the win is convenience. A digital consultant can narrow options, compare textures, and suggest routines that are realistic rather than aspirational. That can make self-care feel more doable on a weeknight, which is one reason beauty continues to behave like a resilient category even when budgets get tight.

7) The Business Case Behind the Experience

AI improves conversion, retention, and media value

From a retail perspective, AI in beauty is not only about novelty. It can improve conversion by helping shoppers choose faster, increase retention by making repeat purchases easier, and raise basket size by surfacing complementary items. It also creates better data for personalization, which can make loyalty programs more effective over time. This is especially important in a category where repeat purchase behavior is a major revenue engine.

There is also a media component. When a retailer owns the customer journey, it can create more relevant on-site placements and product discovery moments. But that only works if the shopper trusts the system. A better recommendation engine is not just a sales tool; it is a relationship builder.

Beauty remains resilient because it feels essential

Circana data cited in the source coverage showed continued growth in prestige and mass beauty, with fragrance and skinification helping drive demand. That matters because it shows consumers still see beauty as emotionally valuable even under affordability pressure. People may trade down on some items, but they continue to invest in products that support confidence, routine, and self-expression.

That resilience makes beauty a fertile ground for tech investment. If the category is already sticky, better guidance can deepen loyalty. The same strategic pattern appears in other categories where perceived value remains strong, such as eco-friendly fashion or curated fragrance edits.

Why Ulta’s scale matters for AI economics

Scale matters because AI systems become more useful when they learn from more interactions, more product attributes, and more customer outcomes. Ulta’s large loyalty base and broad assortment create a strong environment for model training and refinement. That means its recommendations can become more precise over time, especially if the company keeps improving data cleanliness and customer consent infrastructure.

For startups, the challenge is different. They may be more agile, but they often have less data and fewer distribution points. The most successful ones will likely partner with retailers, brands, or platforms rather than trying to own every layer themselves. That is why the future likely includes both large retail AI ecosystems and focused specialist tools.

8) What to Watch Next in Beauty Tech

Expect more agentic shopping flows

The next generation of AI beauty advisors will likely move from recommendation to action. Instead of simply saying, “Here are three moisturizers,” the system may build a regimen, reorder a favorite cleanser when you are running low, and suggest a backup product if your usual item is unavailable. That is the essence of agentic AI: it helps move the shopper from intent to outcome.

That shift could make shopping less fragmented and more helpful. It may also force retailers to think harder about how much autonomy they want the system to have. Helpful is good; overly aggressive automation is not.

Retailers will compete on trust and utility

In the end, the winning digital beauty advisor will be the one shoppers actually use again. That means the interface must be accurate, fast, inclusive, transparent, and respectful of privacy. It also means the system should be useful for both everyday essentials and aspirational purchases. If it can help someone choose a sunscreen today and a fragrance gift tomorrow, it becomes part of the customer’s routine.

For readers who want to think like a smart shopper, the takeaway is simple: let technology narrow the field, but insist on clarity. Ask what data is being used, what the recommendation is based on, and whether the product truly matches your needs. The more informed the shopper, the better the AI can perform.

Beauty shopping is becoming a service layer

What Ulta and the leading startups are building is not just ecommerce. They are building a service layer on top of product assortment—a digital consultant that remembers preferences, interprets needs, and guides decisions. That is a major change from the old beauty counter model, where the best help depended on who happened to be working that day.

The future beauty counter is always open, increasingly personalized, and increasingly data-driven. If retailers get it right, shoppers get better matches, less waste, and more confidence. If they get it wrong, the experience becomes noisy and intrusive. The opportunity is enormous, but so is the responsibility.

Pro Tip: When using any AI beauty advisor, compare its advice with your own skin history, ingredient sensitivities, and return policy. The best results come from AI plus self-knowledge, not AI alone.

Frequently Asked Questions

Conclusion: The Beauty Counter Is Going Digital

Ulta’s AI strategy shows that the beauty counter is no longer confined to a store aisle. It now lives in loyalty data, mobile apps, virtual try-on tools, and conversational interfaces that can guide shoppers from “I have no idea what I need” to “this is the right product for me.” Startups are pushing the category further by making these tools more specialized, more conversational, and more visually intelligent. Together, they are reshaping beauty from a product search into a guided decision experience.

For everyday shoppers, that is good news if the technology is implemented responsibly. The best beauty tech should save time, improve confidence, and reduce waste—not create more noise. As the category matures, the winners will be the brands that combine personalization with trust, and the shoppers who know how to use the tools wisely. For more smart shopping context, you may also enjoy our coverage of retention after the sale, value bundles, and bargain detection.

Related Reading

• How to Use Local Data to Choose the Right Repair Pro Before You Call - A practical look at how trusted data improves buyer confidence.
• How to Build a Governance Layer for AI Tools Before Your Team Adopts Them - A useful framework for responsible AI deployment.
• How to Build an Airtight Consent Workflow for AI That Reads Medical Records - A strong model for consent-first AI design.
• Harnessing AI-Driven Order Management for Fulfillment Efficiency - See how automation improves operational accuracy.
• Airport Fragrance Edit: What to Buy (and Skip) at Goa’s New Olfactive - A curated take on fragrance shopping with a smarter lens.