Every few weeks, CheckoutSmart uses its AI Consumer Panel to explore a new shopper topic and understand what consumers are really asking AI.
Here are the four big takeaways from our latest research:
AI is becoming a new digital shelf. Shoppers are increasingly using AI to shortlist brands, compare options, and decide what deserves attention before they ever reach a retailer site.
The shopper question has changed. In sunscreen and laundry, people are asking about missions such as sensitive skin, SPF, stains, fragrance, babies, ingredients, and product formats rather than simply asking which brand is best.
Only a few brands make the answer. AI responses typically surface a small number of recommendations, meaning brands that are not mentioned may be excluded from consideration before shoppers get close to purchase.
Answer-ready content is becoming a competitive advantage. Brands need clear claims, credible evidence, fresh reviews, strong retailer content, and useful brand-site information that directly answers the questions shoppers are asking.
The goal of the CheckoutSmart AI Consumer Panel is to understand how real consumers are using AI chatbots and AI search tools in everyday decision-making, and to see what answers they receive in return.
This time, we looked at sunscreen and laundry. On the surface, they appear to be very different categories. One sits within beauty and personal care, the other within household products. Yet the shopper behavior was remarkably similar.
In both categories, consumers were not simply asking for brand names. They were asking practical, detailed, purchase-shaping questions about their needs, routines, problems, and priorities.
AI search and large language models are already changing what shoppers see first. They are no longer simply returning links. They are summarizing choices, ranking brands, surfacing retailers, explaining trade-offs, and deciding which product messages are worth repeating.
For CPG brands, this creates a new challenge.
It is no longer enough to be visible in search; brands increasingly need to be answerable.
AI and LLM chatbots are already becoming part of everyday consumer behavior, with around 105 million people in the US regularly using chat bots and younger consumers helping to drive adoption. As Google continues to expand AI-powered search experiences, AI-assisted discovery is becoming increasingly embedded in how people research products and brands.
This matters for CPG brands because Health & Beauty and Grocery are among the categories where consumers are already using AI to ask follow-up questions and refine purchase decisions.
The shift is not simply from search box to chatbot, it is from search results to interpreted answers.
Traditional search engines present shoppers with a list of links and leave them to decide what is most relevant. AI systems increasingly perform part of that decision-making process on the shopper’s behalf. They interpret the question, identify relevant information, summarize options, and often narrow consideration to a smaller set of brands, products, or retailers.
That means the question for brands is changing from:
“Can shoppers find us?”
to:
“Can AI confidently recommend us?”
As AI becomes a more influential part of the path to purchase, brands are no longer competing solely for visibility. They are competing for inclusion in the answer itself.
heckoutSmart’s AI Consumer Panel is designed to show what shoppers are actually asking LLMs, what answers they receive, which brands are being surfaced, and which retailers or sources are referenced in those answers.
The panel consists of tens of thousands of ordinary consumers who actively use AI chatbots and AI-powered search tools in real life. Participants share actual questions and answers across a wide range of topics, including usage behaviours, prompts, responses, attitudes, expectations, and purchase-related decision-making.
For these specific sunscreen and laundry studies, the analysis was based on 727 qualified sunscreen prompt-response pairs and 658 qualified laundry prompt-response pairs after quality control.
The participating sample included a broad consumer mix and was approximately 65% female and 35% male.
That distinction is important. This is not AI asking AI what consumers might ask. It is not a synthetic collection of invented prompts generated by researchers.
It is based on real consumers using AI tools in their everyday lives and sharing the questions they asked and the answers they received.
Real people. Real prompts. Real AI answers.
The result is a more authentic view of AI-influenced shopper behaviour and a clearer understanding of how AI is shaping category discovery, brand consideration, and purchase decisions.
For this research, we asked shoppers what they would ask an LLM about two categories: sunscreen and laundry.
We then collected both the questions and the full answers they received.
This approach provides a much richer perspective than search rankings alone. It allows us to understand the language consumers use, the concerns that matter most to them, the brands AI chooses to surface, the retailers it references, and the claims or product messages that repeatedly appear in answers.
After quality control, the sunscreen study included 727 prompts and responses.
The laundry study included 658 prompts and responses.
For sunscreen, ChatGPT was the most commonly used AI tool (67%), followed by Google LLMs (20%) and other AI tools (13%).
For laundry, ChatGPT represented 59.6% of responses, Google LLMs accounted for 30.2%, and other tools represented 10.2%.
The important point is not simply which AI tool consumers used. It is what their behavior revealed.
Consumers were not using AI as a brand directory, instead they were using it as an advice layer.
That behavior has significant implications for brands because it means AI is increasingly influencing the earliest stages of consideration.
The sunscreen and laundry research shows the same underlying behavior.
In sunscreen, shoppers asked about SPF levels, ingredients, mineral vs. chemical formulas, sensitive skin, makeup, reapplication, cloudy days, winter use, and whether sunscreen should be worn indoors.
In laundry, shoppers asked about tough stains, sensitive skin, allergies, babies, eczema, fresh-smelling clothes, bright whites, fading, ingredients to avoid, and whether to use liquid, powder, pods, sheets, or beads.
The categories are very different, but the shopper behavior is similar. People are not just asking AI for a brand list. They are asking for help with a specific situation.
They want to know what is best for their skin, their family, their clothes, their routine, their budget, or their concern. AI then turns those needs into recommendations.
That means AI visibility is increasingly mission-led. A brand may be strong overall but weaker for a specific need. Another brand may not dominate the category but may win in a particular question type.
In sunscreen, AI leaned strongly toward skincare, facial SPF, protection, and daily use, with facial-SPF brands appearing more prominently than traditional beach sunscreens. In laundry, AI organized answers around practical problems such as stains, fragrance, sensitive skin, and fabric care.
This is one of the biggest strategic lessons: shoppers do not always enter a category through the same door that brands expect.
Takeaway: If your brand content is organized only around your internal brand architecture, you may miss the way shoppers actually ask AI for help. Content needs to match real shopper missions.
One of the most important findings across both categories is how few brands appear in a typical AI answer.
In sunscreen, ChatGPT averaged 5.4 brand mentions per answer, Google averaged 4.7, and other LLMs averaged 3.8. That means most shoppers are seeing only a handful of brands in response to their questions. This is a very different dynamic from traditional search, where dozens of brands may be visible across a results page.
The sunscreen research also showed that brand visibility was concentrated among a relatively small group of players, with skincare-focused facial SPF brands appearing particularly frequently.
The same pattern emerged in laundry. Tide appeared in 55.5% of brand answers, all in 48.7%, and Persil in 29.8%. However, leadership changed depending on the shopper mission. OxiClean performed strongly for stain-removal questions.
Gain was prominent in fragrance-related questions. All appeared consistently across a wide range of sensitive-skin and family-oriented discussions.
This highlights a crucial difference between traditional search visibility and AI visibility. Search visibility was often about appearing somewhere on the page. AI visibility is about being selected, summarized, and recommended. The stakes are therefore much higher.
A shopper may not compare dozens of products or click through multiple websites. They may simply read one AI-generated answer and move forward with the small set of brands included in that response.
For brands, this creates a new form of competitive pressure. The key question is no longer whether your brand ranks somewhere. The question is whether your brand makes the shortlist.
Takeaway: AI visibility is not one ranking. It is mission by mission, question by question. Brand managers need to understand where they are winning, where they are absent, and which competitors AI is recommending instead.
Across sunscreen and laundry, the questions shoppers asked were detailed and practical. That matters because AI systems need clear, structured, credible content to answer those questions well.
For sunscreen brands, that means building content around use cases such as:
• Sensitive skin
• Acne-prone skin
• Oily skin
• Dry skin
• No white cast
• Use under makeup
• Kids
• Sports
• Reef-safe options
It also means making substantiated claims easy to find and understand, including broad spectrum, SPF 30, SPF 50, water resistant, non-comedogenic, fragrance-free, zinc oxide, and dermatologist-tested.
For laundry brands, the same principle applies. Content needs to answer questions about:
• Stains
• Odors
• Sensitive skin
• Babies
• Allergies
• Whites
• Colors
• Fabric care
• Product formats
• Ingredients
Generic product copy is not enough. Brands need content that directly matches the language shoppers use.
The research highlights the importance of content that answers specific comparison and usage questions. For sunscreen brands, that includes topics such as “mineral vs. chemical sunscreen,” “SPF 30 vs. SPF 50,” “face vs. body sunscreen,” and “best sunscreen for dark skin.” These are examples of the types of questions shoppers are asking and the types of content brands should consider addressing.
Takeaway: Brands should stop thinking of content as a static product description. It is now part of the recommendation system. The clearer the answer, the easier it is for AI to understand, summarize, and use.
The research shows that AI answers tend to favour content that is clear, specific, well-structured, and supported by recognisable evidence or authority cues.
That means bold claims are not enough. A brand saying it has the “best” product is less useful than content that explains why a product is suitable for a particular need and supports that explanation with reviews, tests, comparisons, expert endorsement, certifications, or clear product data.
This is especially important in categories where shoppers are asking advice-led questions.
In sunscreen, AI needs to understand why a product may be suitable for sensitive skin, daily facial use, darker skin tones, children, sports, or use under makeup.
In laundry, it needs to understand why a product may be suitable for tough stains, sensitive skin, babies, odour control, whites, or fabric care.
In both cases, AI is not only repeating claims. It is often summarising information that appears credible, useful, and easy to explain.
That authority could come from structured product information, brand websites, retailer content, reviews, FAQs, expert content, testing, awards, or other credible sources. The more clearly that evidence is presented, the easier it becomes for AI systems to understand and repeat.
Takeaway: If a claim matters commercially, it needs support. “Best,” “gentle,” “effective,” “safe,” “dermatologist-tested,” “tough on stains,” or “suitable for sensitive skin” should be backed by evidence that both shoppers and AI systems can understand.
Another important difference between traditional search and AI-led discovery is personalization.
The same question can produce different answers depending on the platform, model version, prompt wording, available search context, location, and, where permitted, user-level signals.
That means brands cannot simply ask one AI tool one question and assume they understand what shoppers are seeing.
One person asking about “best sunscreen” may receive an answer focused on facial SPF and sensitive skin. Another may receive an answer focused on sports, kids, darker skin tones, mineral formulas, or budget options.
The same is true in laundry. “Best detergent” could lead to different recommendations depending on whether the shopper’s implied need is stain removal, fragrance, sensitive skin, babies, whites, value, or convenience.
This variability is one reason why single-prompt testing can be misleading. Looking at one AI answer may reveal what one shopper sees, but it does not reveal the full range of answers that consumers are receiving across different contexts and needs.
This is why real panel-based data matters. To understand AI visibility properly, brands need to see the breadth of questions and the breadth of answers shoppers are actually receiving.
Takeaway: One prompt is not a strategy. Brands need to understand AI visibility across a real range of shoppers, questions, needs, and answer variations.
AI answers do not only refer to brand websites. Retailers are also part of the information ecosystem.
In sunscreen, the retailer data shows Walmart and Target appearing strongly, alongside beauty-focused retailers such as Ulta and Sephora. In the laundry category, Target and Walmart were the strongest retail sites referenced, followed by CVS, Costco, Home Depot, and Amazon.
Amazon was comparatively lower in the early-stage laundry answers, but that does not necessarily mean underperformance. The research suggests Amazon tends to become more relevant later in the purchase journey, when the question is more explicitly about where to buy, price, reviews, delivery, Subscribe & Save, Prime, bulk packs, or marketplace comparison.
This is a critical point for brands. Retailer content is not just about conversion on the retailer page. It can also influence what AI finds, understands, and repeats.
Product titles, descriptions, claims, taxonomy, reviews, availability, and comparison information all matter. If retailer content is thin, outdated, inconsistent, or hard to interpret, AI systems may be less likely to use it effectively.
Retailer choice also reflects the way LLMs try to be useful. If an AI system does not know enough about the shopper, it is more likely to recommend broadly available retailers. That helps explain why retailers such as Walmart and Target can feature strongly in early-stage category answers.
Takeaway: Retailer pages are no longer just product detail pages. They are potential training and retrieval sources for AI-led discovery. Brand managers should treat priority retailer content as part of their AI visibility strategy.
For several years, many brands have focused heavily on retailer execution and paid media, while brand websites have sometimes been treated as lower priority.
AI discovery changes that.
Brand websites can become important sources of authority. They are places where brands can explain the category, answer shopper questions, provide comparison content, publish FAQs, support claims, host reviews, and create useful short articles around specific shopper needs.
That does not mean the brand site replaces the retailer page. Both matter. But the brand site is where a brand has the greatest control over how its products, claims, category expertise, and evidence are presented.
If shoppers are asking AI about sensitive skin, mineral sunscreen, stain removal, fragrance, babies, eczema, or ingredient concerns, brands should have clear, useful, credible answers available in places AI systems can read.
Importantly, this content needs to be organized in ways that AI can easily interpret. Clear FAQs, structured product information, comparison pages, ingredient explanations, and use-case content all make it easier for AI systems to understand when a product should be recommended and why.
The opportunity is not just to describe the product. It is to help AI understand when and why the product should be recommended.
Takeaway: Your brand site should not only sell your brand. It should answer the questions that make your brand recommendable.
Across the research, several factors appear to matter when influencing LLMs: logic and supporting evidence, structured data, fresh information, pre-packaged answers to specific questions, and personalized responses.
In simple terms, AI systems appear to value content that combines relevance, credibility, quality, consensus, and freshness. They also tend to favor information supported by evidence, structured in clear ways, and regularly updated over time.
That is very close to what we already know about shopper reviews. Fresh, credible, representative content helps shoppers make decisions. Now, the same principle is becoming important for AI discovery.
Reviews help because they provide real consumer language. They show how people describe benefits, problems, trade-offs, and use occasions. That language can help AI understand where a product fits and when it should be recommended.
Freshness matters too. Old content and old reviews are less useful. AI systems are more likely to value information that appears current, especially in categories where products, claims, availability, and consumer expectations change.
For brands, this creates a clear link between shopper content, ecommerce content, and AI visibility. The better your content reflects real shopper needs, the more useful it becomes in an AI-led discovery journey.
For many brands, this means AI visibility is becoming closely connected to reputation management. The quality and recency of reviews, product information, retailer content, and supporting evidence increasingly influence not only what shoppers see, but what AI chooses to repeat.
Takeaway: Fresh reviews, current claims, and up-to-date retailer content are no longer only conversion tools. They are also signals that can help AI understand which brands are credible and relevant now.
1. Shoppers ask about missions, not just brands
In sunscreen and laundry, people ask AI for help with specific needs such as sensitive skin, SPF, stains, fragrance, babies, ingredients, and product formats. AI increasingly organizes recommendations around shopper missions rather than broad category leadership.
2. The AI recommendation window is narrow
AI answers usually surface only a small number of brands, so being left out can mean being excluded from consideration before the shopper reaches a retailer.
3. AI rewards specific, answerable content
Generic product descriptions are not enough. Brands need content that directly answers the practical questions shoppers ask, supported by clear explanations, comparisons, and use-case guidance.
4. LLMs Prefer Clear, Evidence-led Content
Claims need support. AI is more likely to use and repeat information that is backed by clear evidence, reviews, testing, expert input, or structured product data.
5. Personalization means there is no single AI answer
Different shoppers can receive different answers to the same question, depending on their needs, context, location, shopping behavior, or platform signals.
6. Retailers are part of the answer too
Retailer product pages, taxonomy, claims, reviews, and availability can all influence what AI finds, trusts, and repeats.
7. Brand websites matter again
Brand sites can act as important sources of authority when they answer real shopper questions through FAQs, comparison content, claims, reviews, and useful category guidance.
8. Reviews and fresh content matter more than ever
Fresh reviews, current claims, and up-to-date product information help both shoppers and AI systems understand which brands are credible, relevant, and worth recommending.
The research points to six practical actions for CPG brand managers.
Do not rely on assumptions or AI-generated prompt lists. Find out what real consumers are asking AI in your category, how often they ask those questions, and what answers they receive. This gives brand, ecommerce, shopper, and insight teams a clear view of where shopper demand is forming.
AI visibility is not one single ranking. It changes by shopper mission.
For sunscreen, priority missions might include sensitive skin, makeup compatibility, SPF confusion, mineral vs. chemical formulas, daily facial use, dark skin, children, or sports.
For laundry, priority missions might include stains, fragrance, sensitive skin, babies, whites, colors, ingredient concerns, or product formats.
Brands need to know which missions they already win, which they are missing, and which competitors are being surfaced instead.
Generic product copy is not enough. AI needs content that clearly answers the way shoppers ask.
That means building FAQs, product pages, comparison content, claim explanations, short articles, and retailer content around real questions such as:
• “What is best for sensitive skin?”
• “Is SPF 50 better than SPF 30?”
• “What removes tough stains?”
• “Are pods, liquid, powder, or sheets better?”
• “Which ingredients should I avoid?”
The clearer and more specific the answer, the easier it is for AI to understand, summarize, and recommend.
LLMs tend to favor claims they can justify. Brands should make sure important claims are backed by clear evidence, such as product testing, certifications, reviews, expert input, awards, ingredient explanations, or usage guidance.
Claims such as “dermatologist-tested,” “fragrance-free,” “non-comedogenic,” “water resistant,” “sensitive skin,” “tough on stains,” or “safe for babies” should be easy to find, easy to verify, and consistently expressed across brand and retailer content.
Retailer pages are not just conversion points. They can influence what AI finds, trusts, and repeats.
Brands should identify which retailers appear most often in AI answers for their category and make sure those product pages are complete, current, well-structured, and rich in useful reviews.
If Walmart, Target, Ulta, Sephora, CVS, Costco, Home Depot, or Amazon are being referenced in AI answers, the quality of content on those platforms matters.
AI-led discovery is not a one-time optimization exercise. Shopper questions change, competitors update their content, retailer pages evolve, and AI systems continue to develop.
Brands should keep claims, product information, FAQs, retailer pages, and reviews current. Fresh, credible content is more useful to shoppers and more likely to be valued by AI systems.
At CheckoutSmart, our AI Consumer Panel helps brands see what shoppers are really asking and what AI is really showing them.
We can help your team understand:
• Which questions shoppers are asking AI in your category
• Which brands are being recommended
• Which claims and messages are being repeated
• Which retailers and other key sources are influencing the answers
• Where your brand is visible, missing, or vulnerable
• What content actions you should take next
If you want to understand how AI is shaping shopper decisions in your category, contact CheckoutSmart to get your own survey.
NB Article originally written in US English