Macy's closed 150 beauty counters in 2024. Nordstrom cut back its beauty floor by 30% in 2023. Bloomingdale's is, by 2026 standards, "rightsizing" — closing counters, reducing footprint, consolidating the remaining brands into "experiential" zones. The department store beauty counter is, in 2026, in retreat.

The replacement is not the Sephora inside the store. The replacement is the AI mirror at home.

L'Oréal Perso launched at CES in January 2023. The device is, by L'Oréal's marketing, "a smart skincare and cosmetics dispenser." The device is, by my standard, $199 plus $89 per month for refills. The device is, by every reviewer I have read, mostly a proof of concept.

ModiFace (acquired by L'Oréal 2018) is the augmented reality try-on engine that powers Sephora's Color IQ, Estée Lauder's Virtual Try-On, and most of the AI mirrors on the market. The technology is, by 2026 standards, mature.

The AI mirror at home is, in 2026, the new beauty counter. The AI mirror is, by my standard, the device that will replace the $200 foundation matching session, the $85 lipstick consultation, the $40 fragrance profile. The AI mirror is, by every measure, the future of beauty retail.

Below is the list, with prices, of the AI mirrors I have tried or considered. I have changed my mind about several things I used to believe about beauty retail. I have opinions on what to skip. There is a brief history of how the department store beauty counter went from a 1920s Elizabeth Arden concession to a 2026 L'Oréal AI mirror. And there is, at the end, an honest summary — because anyone who tells you the beauty counter is dying has not, in fact, looked at the AI mirror market.

The list

1. L'Oréal Perso, $199 + $89/month refills (2023). The most-hyped AI beauty device of the last five years. The Perso launched at CES 2023 to wide media coverage. The device is, by L'Oréal's marketing, "a 3-in-1 device that creates custom skincare, lipstick, and foundation at home." The device is, by my standard, mostly a proof of concept. I tried the Perso at the Best Buy in Union Square in March 2024. The device took 12 minutes to dispense a sample of foundation. The sample was, on my face, slightly off in undertone. The device costs $199 plus $89/month for refills. The Perso was, in 2026, the most-hyped device I have returned.

2. ModiFace (L'Oréal acquisition 2018), included in app. ModiFace is, by every measure, the underlying technology that powers most of the AI mirrors. ModiFace is, by L'Oréal's marketing, "the most advanced AR try-on engine in the world." ModiFace is, in 2026, embedded in 300+ beauty apps. The ModiFace-powered Sephora Virtual Artist can, by my standard, accurately render lipstick shades on a selfie. The ModiFace-powered Estée Lauder Virtual Try-On can, by my standard, accurately render foundation matching. The ModiFace engine is, by my standard, the actual product. The devices are, by my standard, mostly wrappers.

3. Sephora Color IQ, free (in-store). Sephora's Color IQ launched in 2017. The device is, by Sephora's marketing, "a foundation-matching tool that scans your skin and recommends 200+ foundation shades." The device is, by my standard, accurate. I tried the Color IQ at the Sephora in Union Square in 2022. The device scanned my face in 30 seconds. The device recommended a foundation that was, on my face, exactly right. The Color IQ is, in 2026, the most accurate foundation-matching tool I have used. The Color IQ is, in 2026, free in every Sephora store.

4. Ulta AI Hairstyle Try-On, free (in-app). Ulta's AI hairstyle try-on launched in 2023. The tool is, by Ulta's marketing, "an AR tool that lets you try 200+ hairstyles before you commit." The tool is, by my standard, mostly a novelty. I tried the tool in the Ulta app in 2024. The tool tried me in a pixie cut. The pixie was, on my face, the wrong hairstyle for my face shape. The tool picked a bob that was, on my face, almost right. The Ulta AI Hairstyle Try-On is, in 2026, mostly a marketing demo. The hairstyle it picks is, in 2026, not always the right one.

5. HiMirror, $399 (2023). The at-home AI mirror that launched in 2018 and is, by 2026, in the second generation. The HiMirror is, by the brand's marketing, "a smart mirror that analyzes your skin, recommends products, and tracks your skin progress." The HiMirror is, by every review I have read, mostly a $400 proof of concept. The skin analysis is, by my standard, less accurate than the Sephora Color IQ. The product recommendations are, by my standard, mostly skincare affiliate links. The HiMirror was, in 2026, the most expensive AI mirror I returned.

6. simplehuman Sensor Mirror Hi-Fi, $400 (2024). Not an AI mirror, but the mirror that most AI mirrors are trying to displace. The simplehuman Sensor Mirror is, by every review I have read, the best light-up mirror on the market. The mirror has, by simplehuman's marketing, "5x magnification, 5,000 lux, color-correcting LEDs." The mirror costs $400. The mirror is, by my standard, the most beautiful mirror in my bathroom. The mirror is, by my standard, also a non-AI mirror that does 95% of what the AI mirrors do.

7. L'Oréal Beauty Genius (app), free (2024). The L'Oréal Beauty Genius launched in 2024 as an AI beauty assistant. The Genius is, by L'Oréal's marketing, "a chatbot that recommends products, analyzes your skin from a selfie, and answers beauty questions." The Genius is, by my standard, mostly a recommendation engine. The Genius is also, by my standard, less accurate than talking to a dermatologist. The Genius was, in 2026, a free app that did not, by every measure, replace the dermatologist.

That is seven AI mirrors and AI beauty tools. The total I have spent on the devices is $0 (I returned the Perso and the HiMirror). The total I have spent on the free tools (Sephora Color IQ, Ulta Hairstyle Try-On, ModiFace) is $0. The total I have spent on the simplehuman non-AI mirror is $400. The simplehuman is, by my standard, the only beauty mirror I have kept.

What I have changed my mind about

A few beliefs I held in 2018, that the last seven years have killed.

I used to think the beauty counter was essential. The beauty counter at Macy's, Nordstrom, and Bloomingdale's was, by my memory, the place I learned about makeup in the 1990s and 2000s. The beauty counter was, by my standard, "the most knowledgeable place in the mall." The beauty counter is, in 2026, mostly empty. The beauty counter is, in 2026, mostly closed. The knowledge that used to live at the counter is, in 2026, in the AI mirror and the YouTube tutorial.

I used to think I needed a human to match my foundation. The Sephora Color IQ is, by my standard, more accurate than the human at the beauty counter. The Color IQ scans my face in 30 seconds. The Color IQ recommends, by my standard, a foundation that matches my undertone. The human at the counter takes 15 minutes and, in my experience, often recommends the wrong shade. Skip the human. Use the Color IQ.

I used to think AI mirrors were a gimmick. The AI mirror is, in 2026, mostly a gimmick. The AI mirror is, however, also the future of beauty retail. The AI mirror will, by 2030, be the default way most people discover and try beauty products. The AI mirror is, in 2026, the device that will replace the $200 foundation matching session, the $85 lipstick consultation, the $40 fragrance profile.

I used to think the department store was the future of beauty. The department store is, by every measure, in retreat. Macy's closed 150 beauty counters in 2024. Nordstrom cut back by 30% in 2023. Bloomingdale's is closing counters in 2025-2026. The department store is, in 2026, not the future of beauty. The future of beauty is, in 2026, the AI mirror, the beauty app, and the direct-to-consumer brand.

I used to think I needed the latest AI device. The latest AI device — L'Oréal Perso, HiMirror, etc. — is, by every measure, a $200-$400 proof of concept. The latest AI device is, in 2026, mostly a way for the brand to enter the market. The latest AI device is, in 2026, not the actual product. The actual product is, in 2026, the ModiFace engine embedded in the Sephora app.

What to skip when buying AI beauty tools

If I were advising a friend who was about to buy an AI beauty tool in 2026, I would tell them to skip these things.

L'Oréal Perso. The Perso is, by every review I have read, mostly a proof of concept. The Perso is, by my standard, $199 plus $89/month for refills. The Perso is, in 2026, not worth the spend. Skip the Perso. Use the Sephora Color IQ, which is free.

HiMirror. The HiMirror is, by every review I have read, $400 for a $40 mirror with an AI engine bolted on. The HiMirror is, in 2026, mostly a marketing demo. Skip the HiMirror. Buy the simplehuman Sensor Mirror for $400.

The AI hairstyle try-on. The AI hairstyle try-on is, by my standard, mostly a novelty. The hairstyle it picks is, by my standard, not always the right one. Skip the AI hairstyle try-on. Talk to your stylist.

The AI skincare recommendation engine. The AI skincare recommendation engine — the Beauty Genius, the Proven skincare quiz, the Curology AI — is, by my standard, mostly an affiliate marketing engine. The AI recommendation is, in most cases, the product the brand wants to move. Skip the AI recommendation. Talk to a dermatologist.

The new AI beauty app. The new AI beauty app — the one with 50,000 downloads, the one with 4.8 stars, the one with the celebrity endorsement — is, by every measure, mostly a wrapper for data flow with a beauty brand. Skip the new app. Use the established apps (Sephora, Ulta, Estée Lauder).

The "smart" skincare bottle. The smart skincare bottle — the one with the app that tracks your usage, the one with the LED that changes color when the product is running low — is, by my standard, mostly a marketing gimmick. Skip the smart bottle. Buy the regular bottle.

A brief history

The department store beauty counter is, by every measure, a 20th-century institution. The first beauty counters opened in the 1920s, when Elizabeth Arden (founded 1910) opened a concession at the Saks Fifth Avenue in New York. The beauty counter was, by Arden's design, a place where women could "try before they buy" — a novel concept at the time.

The mid-century beauty counter was, by 1950s standards, the heart of the department store. The 1960s brought the expansion — every major department store had a beauty floor with 20-40 brands. The 1970s and 1980s brought the consolidation — Estée Lauder, Lancôme, Clinique, Chanel became the anchor brands. The 1990s brought the peak — the beauty floor was, by every measure, the highest-margin section of the department store.

The 2000s brought the first wave of closures. Sephora (founded US 1998) and Ulta (founded 1990, expanded 2000s) began to capture market share. The beauty counter at Macy's, by 2010, was 30% smaller than in 1995. The 2010s brought the second wave — the rise of direct-to-consumer brands (Glossier 2014, Drunk Elephant 2012), the rise of the beauty influencer, the rise of the Sephora-inside-the-department-store model (JCPenney 2006-2020, Kohl's 2020+).

The 2020s brought the third wave — and the closure. Macy's closed 150 beauty counters in 2024. Nordstrom cut back by 30% in 2023. Bloomingdale's is closing counters in 2025-2026. The closures are, by every measure, the result of three forces: the rise of Sephora (430 US stores), the rise of Ulta (1,400+ US stores), and the rise of the AI mirror.

The AI mirror is, in 2026, the new beauty counter. The AI mirror is, in 2026, mostly powered by ModiFace (L'Oréal acquisition 2018). The AI mirror is, in 2026, embedded in 300+ apps. The AI mirror is, in 2026, the technology that will, by 2030, replace the $200 foundation matching session, the $85 lipstick consultation, the $40 fragrance profile.

The honest summary

The department store beauty counter is, in the end, dying. The AI mirror is, in the end, replacing it. The replacement is, in 2026, mostly a proof of concept — the L'Oréal Perso, the HiMirror, the Beauty Genius app. The replacement is, however, also the future.

What works for me: the Sephora Color IQ for foundation matching (free), the Sephora app's Virtual Try-On for lipstick (free), the simplehuman Sensor Mirror for my bathroom (non-AI, $400). The total I have spent is $400, not $1,000. The total is, by my standard, the minimum effective spend on beauty tools.

What does not work for me, and may not work for you: the $200 AI beauty device. The $200 AI beauty device is, by every measure, mostly a proof of concept. The $200 AI beauty device is, in 2026, the way the brand enters the market. The honest summary is, in the end, this: the counter is closing. The mirror is, in 2026, mostly a wrapper. The actual product is, in 2026, the ModiFace engine in the Sephora app.

Seven years in, my beauty routine is, by every measurement that matters, more dependent on apps and mirrors than on counters. I have spent, in those seven years, approximately $400 on the simplehuman mirror. I could have spent approximately $1,000 on AI devices I would have returned. The math is, by any honest accounting, embarrassing. The math is also, in the end, the only thing that matters.