AI in Beauty Customer Service: What’s actually happening vs What brands think is happening

Before getting into the data, it is worth defining terms. AI customer service is not one thing, and the industry uses three terms interchangeably, which is part of the problem.

Chatbot 

Rules-based and menu-driven. You select from a list of options, it routes you somewhere. This is what shaped most consumer sentiment around AI customer service, and it has very little to do with what brands are deploying now.

Conversational AI.

Uses natural language processing to understand open-ended questions and respond accordingly. Can handle nuance and interpret intent. Widely used in beauty for product recommendations and skin consultations.

Agentic AI

Where the market has moved. Agentic AI does not just respond, it acts. Processes refunds, initiates returns, modifies subscriptions, updates orders, and escalates to human agents with full context already attached. It connects directly to back-end systems and executes tasks end to end.

The confusion between these categories is not semantic. Brands are reading consumer distrust data about chatbots and applying it to agentic AI investment decisions. That is a category error, and it is driving underinvestment based on the wrong reference point.

The numbers that are shaping brand decisions look alarming on the surface. According to Vogue Business’s April 2026 consumer survey, 55% of beauty consumers distrust AI recommendations, and 66% say their in-store experience would be worse if AI were involved. Brands read those figures and pull back from investment.

That is the wrong response to the right data.

The same survey shows that only 2% of consumers say AI gets their preferences right consistently. But 62% say it gets them right sometimes. Consumers are not saying they do not want AI involved in their experience. They are saying the versions they have encountered so far have not been good enough. Those are very different briefs.

The data privacy picture tells a similar story. 72% of consumers say they would not share card details with an AI system. Presented in isolation, that sounds like a hard no. But 49% are perfectly comfortable sharing their dress size. Consumers are drawing proportional lines around sensitive information, which is entirely reasonable, and it is something brands can design around with thoughtful data collection flows and clear consent architecture.

Worth noting too: 49% of consumers do not trust influencer recommendations either. AI is not uniquely distrusted, it is equally unproven. The difference is that brands have spent years learning how to make influencer partnerships work despite that distrust. The same investment in understanding is warranted here.

The consumer resistance to AI in beauty customer service is real. It is just not the barrier most brands think it is. The signal in the data is that execution quality is everything. Brands that deploy AI well are dissolving this distrust.

Most brands picture conversational AI when they think about automating customer service. A widget that answers product questions, handles basic queries, maybe recommends a moisturiser. That technology exists and works well. It is also no longer the most relevant part of the conversation.

The market has moved to agentic AI. Agentic AI systems do not just respond to queries, they execute actions within connected back-end systems. Processing a refund. Initiating a return. Modifying a subscription. Updating a delivery address. Escalating a complex case to a human agent with the full conversation history already attached. The distinction matters because it changes the commercial case entirely. This is about resolving customer queries, end to end, without human intervention.

The brands already operating at this level are producing results that are difficult to argue with.

Sephora’s virtual try-on and AI consultation tools generated over 200 million shade trials within two years of launch. E-commerce revenue grew from $580 million to over $3 billion between 2016 and 2022, with AI-assisted appointment booking driving an 11% uplift in booking rates. These represent a structural shift in how the brand converts digital traffic.

Nip + Fab’s integration with Renude’s AI skin consultation tool delivered a 16% conversion rate and a 42% uplift in average order value. The detail worth noting: the brand’s CX director tested the tool’s ingredient and formulation knowledge in person before committing. She wasn’t just sold on a concept but on the quality of the output. That is a useful model for how beauty brands should be evaluating AI vendors.

YISE Beauty is arguably the most instructive example for growth-stage brands. Operating across Shopify DTC and Sephora, the brand deployed an AI agent called Dottie, built on StateSet’s iCommerce Engine, early in their scaling curve, before the operational pressure became unmanageable. The results: 46% of incoming tickets handled end to end by AI, a first response time of 1 minute 29 seconds against a merchant average of 13 hours 26 minutes, and a CSAT score of 4.83 out of 5 maintained as ticket volume grew 3x. The CX team saved an estimated 96 hours per month, time that was reallocated to VIP retail launches and product education rather than routine query handling.

YISE Beauty deployed it as infrastructure, early enough that it shaped their growth rather than reacting to it. StateSet‘s system connected directly to Shopify, Recharge, and shipping carriers, meaning Dottie was not just responding to customers but executing actions inside live operational systems from day one. That sequencing and that level of integration matters.

Looking ahead, Gartner projects that by 2029, agentic AI will autonomously resolve up to 80% of common customer service issues without human involvement. The brands building capability now will have accumulated data, refined their models, and established consumer trust in their AI interactions well before the majority of the market catches up.

Efficient customer service is a worthwhile outcome. It is also the smaller part of what AI makes possible in beauty.

The more significant commercial opportunity sits earlier in the purchase journey, at the point where a consumer is trying to decide what to buy and does not yet have the confidence to commit. According to the Revieve Beauty and Wellness Index 2025, 24% of beauty consumers do not know their skin type. One in four shoppers arrive on a beauty site without the foundational knowledge to make a confident purchase decision. Every one of those shoppers is a potential return, a missed basket, or a lost customer who clicked away rather than risk getting it wrong.

AI-guided diagnostic tools change that dynamic. Where traditional ecommerce asks consumers to do the work, guided discovery reverses that logic. Instead of presenting a catalogue and expecting shoppers to self-select, diagnostic tools ask questions, interpret answers, and surface the right products for that specific person’s needs. The conversion impact is material: consumers using guided diagnostics converted at up to 1.9x higher rates during the 2025 peak trading period. Between 60% and 75% of beauty shoppers now begin their purchase journey with some form of guided diagnostic experience.

The implication for how brands should be categorising this investment is significant. Guided commerce powered by AI is a revenue driver. Nip + Fab’s 42% uplift in average order value through Renude’s consultation tool makes that case clear. Consumers who are guided to the right product spend more, return less, and come back.

For premium and high-science beauty brands in particular, this represents a structural advantage. The more complex the formulation, the more a consumer benefits from being guided rather than left to browse. AI consultation scales that expertise without scaling headcount, making it accessible at every hour and every price point across the range.

Beauty discovery happens on social. Instagram has 2 billion monthly active users. TikTok sold approximately 370 million beauty and personal care products directly on the platform last year. And yet almost no beauty brands have meaningful AI customer service presence on either channel.

Fashion and travel have already moved. H&M, ASOS, and KLM are operating AI-assisted customer interactions through social and messaging channels. Beauty has been slower, partly due to platform limitations, and partly due to a tendency to treat chatbots as a website tool rather than a channel-agnostic capability. Instagram DMs are where to watch. The Messaging API is opening up, and the brands that build capability there early will have a meaningful head start.