AI Captures 37% of Beauty Searches, Ditching Google

37% of beauty consumers use AI like ChatGPT for personalized product searches, abandoning Google (80% drop-off); brands must weigh owned personalization tools against AI optimization to capture traffic and sales in $450B industry.

AI's Rapid Takeover in Personalized Beauty Searches

Beauty shoppers crave specificity—skin types, curl patterns, aging concerns—that generic Google results fail to deliver, leading 80% to abandon traditional searches. Instead, 37% now turn to AI platforms like ChatGPT, Gemini, Perplexity, and Claude for tailored advice, with 27% of UK shoppers completing purchases via AI agents. This shift hits hardest in a $450B recession-resilient sector where average UK spending rose from £291 (2024) to £324 (2025), fueled by the "lipstick effect" of consistent beauty buys amid economic uncertainty.

Charlie Marchant, CEO of Exposure Ninja, attributes this to longstanding personalization needs: "Personalization is huge in this sector... generic advice is generally not ideal." Smart brands like The Ordinary (skincare regimen builder) and Only Curls (curl-type quiz) preempted AI by creating owned tools that quiz users on concerns like acne or dryness, funneling them to products and enabling segmented email campaigns (e.g., SPF reminders). AI elevates this—users upload skin photos for instant regimens—making it the "next level" for research-to-purchase journeys.

Quote: "ChatGPT is basically my Bible when it comes to everything in this sector." – Charlie Marchant, explaining his personal reliance on AI for color analysis and recommendations, highlighting consumer trust in AI for top-of-funnel discovery.

Budget Tradeoffs: Owned Tools vs. AI Recommendations

Marketing leaders face a core dilemma: invest in proprietary regimen builders (guaranteed brand exposure, repeat revenue via emails/subscriptions) or optimize for AI mentions (top-of-funnel reach to 37% of searchers, but no control over rankings)? Marchant advises both ideally, but context matters.

For large catalogs (e.g., The Ordinary's hundreds of SKUs), owned quizzes reduce overwhelm, segment users, and drive subscriptions. Mid-market brands with 5-6 hero products skip this—focus on clear navigation and AI visibility instead. Competing against giants? Prioritize niche AI optimization: structure sites for easy crawling, earn third-party PR for authority.

Tradeoffs are stark. Owned tools build loyalty but demand dev resources and traffic to shine. AI optimization yields referral traffic (ChatGPT/Gemini links) but risks zero mentions. "The ultimate answer is you want to do all of it," Marchant says, but smaller players win by niching down: "You're more likely... to do well optimizing for AI searches... for the niche of products that you actually have."

Quote: "If you're not in the recommendations, right? You're not pushing any traffic through from those sites... Quite a large chunk of the market to be ignoring." – Charlie Marchant, underscoring the funnel risk of skipping AI, where 37% start their journey.

AI Overviews Reshape Google Traffic Dynamics

Google's AI Overviews appear on 36% of beauty queries, often above organic results, prioritizing problem-solving (e.g., product comparisons, compatibility). 45% of sector marketers report traffic gains post-inclusion, countering SEO debates—especially with visuals boosting brand stickiness.

Optimization diverges from traditional SEO: target high-AI-overview keywords (top/mid-funnel with commercial intent), leverage existing top-10 organics (boosts inclusion odds). Unlike agentic AI (ChatGPT), Overviews demand structured data, authority signals. Marchant: Prioritize where you're competitive; it's a "different project" from chatbot rankings.

Omni-Channel Gaps: 29% Hit Stores for Missing Online Details

Online research influences in-store buys, yet 20% skip e-comm due to confusing descriptions, and 29% enter physical stores (Boots/Superdrug) just to read labels. This inefficiency—navigating overwhelming shelves—stems from poor online clarity.

Fix: Crystal-clear descriptions answering usage, compatibility, concerns. AI/SEO feeds this research phase; omni-channel wins when online informs offline (e.g., avoiding shipping for one-offs). All tactics—AI, Overviews, quizzes—interconnect: "They all actually influence what gets purchased both online and offline."

Quote: "Imagine going into a store... to read the label on the bottle... What a waste of time." – Charlie Marchant, spotlighting the low-hanging fruit of better online descriptions to convert researchers instantly.

Agentic Future Demands Structural Shifts

Projections: Agentic AI (autonomous buyers) hits 10-20% of US e-comm by 2030. Beauty brands restructure now—PR for authority, analytics tracking, category prioritization (e.g., hero SKUs). Track AI referrals in Google Analytics; pull third-party mentions (biggest untapped lever).

Quote: "AI is kind of the next level of personalization... now we can... take a photo of our skin and upload it." – Charlie Marchant, framing AI as evolution from quizzes, urging brands to adapt before competitors.

Key Takeaways

  • Download Exposure Ninja's Beauty AI Search Report for full data and tactics: https://exposureninja.com/beauty-ai-search-report/.
  • Build owned personalization (quizzes/regimens) if you have large catalogs; niche AI optimization otherwise.
  • Target AI Overviews on 36% of queries via top-10 organics and structured content for 45% potential traffic lift.
  • Fix product descriptions to cut 29% store-only research trips and boost immediate online/in-store conversions.
  • Track AI referrals in Google Analytics; prioritize PR for recommendations amid 37% adoption.
  • Weigh business stage: acquisition via AI for traffic-starved sites; retention via emails for high-traffic ones.
  • Prepare for 10-20% agentic purchases by 2030—structure for authority and omni-channel influence.
  • Use tools like Semrush for keyword/AI analysis; test skin photo uploads in ChatGPT for consumer insights.

Summarized by x-ai/grok-4.1-fast via openrouter

8794 input / 2225 output tokens in 20378ms

© 2026 Edge