73% of Consumers Ask AI to Compare Products. What Should Brands Do?

Comparing options is one of the main purposes of using AI for shopping. However, consumer behavior does not stop there.

Macromill identified 12 purposes of AI use among the 882 respondents who use AI for shopping or product research. Respondents could choose more than one answer. When all the percentages are added up, the total reaches 725%, equivalent to an average of about seven purposes chosen per respondent.

This figure shows that consumers use AI for several needs at once during the consideration process. To understand the implications for brands, the 12 purposes can be grouped into three main jobs.

1. Comparing options

73% of respondents use AI to compare several options or technical specifications. 65% compare prices, and 53% compare features.

At this stage, consumers need information that lets them see the differences between products directly. Specifications, prices, features, and variants serve as material for narrowing down choices.

2. Matching products to needs

66% of respondents look for products for specific needs, while 63% look for more personalized recommendations.

This means consumers do not only want to know which product has the best specifications. They also want to understand which product best suits their own circumstances and needs.

The product with the most features or the highest price is not necessarily the most relevant choice for every consumer. Fit with needs is an important part of the evaluation.

3. Verifying claims and credibility

65% of respondents use AI to check a brand’s credibility or trustworthiness, 63% to assess product effectiveness, and 55% to check ingredients or composition. Another 55% also use AI to summarize reviews.

Interestingly, looking for promos or the best deals ranks last among the 12 purposes, at 43% of respondents.

For brands, this pattern shows that using AI for shopping goes beyond searching for prices or recommendations. Consumers also use it to assess products and check information before making a decision.

The implication is that a strategy for making a brand appear in AI answers needs to be followed by efforts to ensure that product information is complete enough to be compared, matched to needs, and verified.

The information consumers need is not always the same. The differences by income, age, and gender in the Macromill survey show that brands need to understand the context behind prospective buyers’ questions, rather than simply offering one kind of message for everyone.

58% of respondents actively ask AI specific questions, 38% ask general questions and then explore the results, and about 4% read recommendations passively or rarely interact.

In the income group above Rp8.5 million, the proportion of respondents who ask specific questions reaches 66%. This is significantly higher than in the Rp3.5 million–Rp8.5 million income group, where it reaches 48%.

As consumers’ questions become more specific, the product information available needs to be detailed enough to answer them. Descriptions that contain only general claims risk failing to adequately answer questions about specifications, fit, or differences between variants.

Income affects the kinds of questions asked

Differences in needs are also visible in the purposes for which AI is used.

The Rp3.5 million–Rp8.5 million income group is the most likely to use AI to compare options, at 82%. This figure is 35 percentage points higher than for the group earning below Rp3.5 million, and Macromill flags the difference as significant.

Meanwhile, the group earning below Rp3.5 million leans more toward using AI to find solutions to problems and more affordable alternatives. 70% of this group look for cheaper alternatives.

The base for the below-Rp3.5 million income group is relatively small, at 175 respondents. Differences that are not flagged as significant should therefore be treated as indications rather than conclusions that apply to all consumers in the group.

For brands, these findings point in two directions for preparing product information. First, provide information that makes it easier for consumers to compare options. Second, explain alternatives by price, size, or variant so that consumers can judge which option fits their budget.

Age and gender also differentiate needs

Respondents aged 17–29 most often look for products for specific needs, at 76%. By contrast, respondents over 40 are more likely to look for personalized recommendations (71%) and to check brand credibility (68%).

Male respondents also use AI to compare features more than female respondents, at 62% and 42% respectively. The Macromill report flags this gap as significant.

These differences do not mean that every consumer in a segment behaves the same way. However, the findings show that information needs can differ across groups. Brands need to test whether their product content can answer the various questions relevant to their category and consumer segments.

Consumers use more than one AI platform

The search journey also takes place across multiple platforms. In the three months before shopping, ChatGPT was used by 70% of respondents, followed by Gemini (58%), Meta AI on Instagram and WhatsApp (40%), AI Mode or the Google assistant (29%), and TikTok’s AI and recommendation features (19%).

Meta AI usage reaches 45% in the 30–39 age group and 49% among respondents over 40, compared with 22% in the 17–29 age group.

For brands, this finding shows the importance of checking information across multiple platforms, including AI assistants integrated into messaging and social media apps. However, being present on many platforms is not enough. The information delivered must also be accurate and consistent when consumers move between channels.

Product Information That Is Useful to AI Has Three Characteristics

If consumers use AI to compare, match, and verify products, the next question is: what kind of information do brands need to prepare?

Based on the three consumer jobs, there are three characteristics of information that are relevant as a basis for evaluating product content.

1. Comparable: enabling products to be compared directly

Product information needs to use clear specifications, units, terminology, prices, variants, and terms so that it can be placed side by side with other products.

When 73% of respondents compare options or technical specifications, 65% compare prices, and 53% compare features, general claims such as “more economical,” “durable,” or “superior performance” need to be accompanied by relevant explanations so that consumers can evaluate them.

For brands, the implication is to present product facts in an explicit and consistent format. Specification tables, price breakdowns by variant, and feature explanations can help answer comparison questions more clearly.

2. Contextual: explaining who the product is suited for

Product information also needs to explain the needs a product can meet, the conditions of use, and the product’s limitations.

This need is consistent with the finding that 66% of respondents look for products for specific needs and 63% look for personalized recommendations. Macromill also noted complaints that AI recommendations are sometimes too generic and the information is not yet up to date.

For brands, product pages should not only describe what the product has, but also help consumers understand when a feature is useful and under what conditions the product may be less suitable.

Contextual explanations help consumers assess fit, rather than simply collecting a list of advantages.

3. Verifiable: providing a basis for verifying claims

Product information needs to have a basis that can be checked. This is relevant to the use of AI to check brand credibility (65%), product effectiveness (63%), ingredients or composition (55%), and to summarize reviews (55%).

The form of evidence depends on the product category. Brands can prepare test results, distribution permits, relevant certifications, ingredient lists, or other documentation that supports claims in accordance with the requirements of the category.

Consistency is also important. If price, composition, or specification information differs between the official website and marketplaces, consumers who double-check may find information that is inconsistent.

Why do brands’ information sources matter?

BCG’s qualitative research on 367 GenAI users in the United States provides additional context. Consumers value AI because it can give direct answers, feels more objective when weighing pros and cons, and helps them articulate what they actually need.

BCG advises brands to prepare explanations in simple language, side-by-side comparisons, and supporting evidence that answer consumers’ real questions.

Meanwhile, the preprint study by Kumar (2026) analyzed more than 100,000 AI answers to questions about more than 100 brands between March and May 2026. About 78% of the sources cited by AI engines came from company websites. Outside company websites, YouTube was cited most often, followed by Reddit, editorial media, and Wikipedia. Ranking-list articles, or “best-of” listicles, accounted for about 21% of all citations.

The study also found that small and niche brands appeared in only about 11% of relevant answers.

These findings should be read with caution. The Kumar study has not been peer-reviewed, its author is affiliated with Ranqo, the visibility-tracking platform that is the source of the data, and its scope is global. These figures cannot be assumed to represent all AI platforms or the behavior of Indonesian consumers.

Even so, the findings provide context for one strategic implication: a brand’s official website is one of the important sources for AI answers. For this reason, the completeness, clarity, and currency of information on the official website deserve priority in an AI visibility strategy.

AI Helps Decisions, but Consumers Still Need to Check

How big a role does AI play in the final purchase decision?

In the Macromill survey, 94% of AI users for shopping say that AI helps their shopping decisions. 51% say AI is very helpful, while 43% say it is fairly helpful.

However, the judgment that AI is helpful does not automatically mean that consumers hand the final decision over to it.

51% of respondents say AI influences or strongly influences their final purchase decision, made up of 33% who say it is influential and 18% who say it is very influential. Another 41% say it is moderately influential.

A very strong influence is seen in the income group above Rp8.5 million. 28% of this group say AI is very influential on their purchase decisions, which is significantly higher than in the Rp3.5 million–Rp8.5 million income group (14%).

There is also a pattern that warrants further investigation. The middle-income group uses AI most for comparing options (82%), but has the lowest proportion rating AI as very helpful (39%), compared with 58% and 62% in the other two income groups.

The survey does not explain the cause of this difference. One hypothesis that could be tested is whether consumers who compare more intensively need more complete information than what is available in AI answers. However, this hypothesis requires additional evidence before any conclusion can be drawn.

When consumers still doubt accuracy

Findings from outside Indonesia show why verification remains an important part of the AI-assisted shopping journey.

A Gartner survey of 846 US consumers in November–December 2025 found that 54% of AI users for shopping feel the need to double-check the accuracy of all the information from AI. 62% say AI information ultimately wastes their time.

In a separate survey of 322 US consumers in January 2026, only 11% were willing to let AI make purchase decisions, even for low-risk categories such as personal care products and household necessities. Willingness to let AI narrow down options was higher, at 31% for household necessities and 28% for personal electronics.

BCG, on the other hand, noted that more than 60% of consumers in nine countries place high trust in GenAI results.

These findings do not necessarily contradict each other. Consumers can find AI useful while still checking the accuracy of the information it provides.

For brands, this means that the quality of AI answers needs attention alongside visibility. When AI states the wrong price, gives inaccurate specifications, or recommends a product that is less suitable, consumers have to make extra effort to check the information. As Kate Muhl, VP Analyst at Gartner, highlighted, accuracy has become a relevant issue for brands.

In other words, an AI visibility strategy should not target only how often the brand name appears. Brands also need to test whether the information that appears is correct, complete, and consistent with the products they offer.

For Brands: Five Steps to Prepare Information for the AI-Assisted Shopping Journey

Generative engine optimization (GEO) is often discussed as an effort to increase the likelihood that a brand appears in AI answers. However, the consumer behavior findings above show that a brand’s work does not end at visibility.

Brands need to prepare information that can answer prospective buyers’ specific questions at the consideration, or pre-purchase, stage. The following five steps can serve as an initial agenda.

1. Audit AI answers using real consumer questions

Test products through comparison, fit, price, and credibility questions in ChatGPT, Gemini, and Meta AI. Check whether the prices, specifications, features, composition, and claims that appear are accurate and consistent with official information.

Do not just record whether the brand is mentioned. Also record incorrect information, unanswered parts, and claims without a clear basis. This approach is in line with BCG’s recommendation to measure visibility together with the quality and accuracy of AI answers.

2. Make the official website the complete source of product information

Ensure that every product page includes specifications, variants, relevant prices, usage instructions, product limitations, and supporting evidence for claims, as appropriate to the category.

Prioritize information that answers consumers’ questions, not just promotional narrative. A product page that is complete, current, and easy to understand is a foundation that should be checked in a GEO strategy.

3. Align information across all channels

Audit the consistency of facts among the official website, marketplaces, product catalogs, and other content channels. Differences in price, composition, variant names, or specifications need to be traced and corrected at the relevant source.

Consistency does not mean that all channels must use the same format. What needs to be maintained is the sameness of the core facts, while explanations can be adapted to the function of each channel.

4. Prepare information according to segment needs

Use the segmentation findings as a basis for testing different information needs. For questions about affordable alternatives, make sure the price per size or variant can be compared. For questions about fit, provide a selection guide based on needs and conditions of use.

Avoid treating the segmentation findings as universal rules. Use the Macromill results to formulate hypotheses, then test their relevance with your own brand’s category and consumers.

5. Build honest and verifiable comparison content

Prepare comparison tables, product selection guides, FAQs, and educational content that explain both the strengths and the limitations of the product. Every claim needs an appropriate basis.

Content that only praises one’s own product does not necessarily answer the questions of consumers who are weighing alternatives. Focus content on consumers’ decision needs: what the differences are, when the product is suitable to use, what to watch out for, and how its claims can be checked.

What should brands monitor?

As a first step, brands can monitor the following three operational indicators:

These indicators are proposed operational measures, not standard metrics validated by the surveys in this article. Brands can set the frequency of checks and the priorities for fixes according to their category, claim risk, and available resources.

From Visibility to Quality of Consideration

The use of AI in shopping opens up new space for brands to be present when consumers compare, look for fit, and check product credibility. However, appearing in AI answers is only one part of that process.

The more strategic question is whether the information a brand holds is strong enough to help consumers evaluate their options properly. Comparable specifications, contextual explanations, and verifiable claims are a foundation that can be put in order without waiting for further changes in AI technology.

Decision question: If a prospective buyer asked AI today to compare your product with its closest competitor, would the answer use accurate, complete, and consistent information about your brand?

Signal to monitor: The accuracy of prices, specifications, and product claims in ChatGPT, Gemini, and Meta AI answers to comparison and fit questions in your category. Record the errors found, trace their sources where possible, and use the audit results to set priorities for content improvements.

References

Gartner. (2026, May 27). Gartner survey finds consumers want AI shopping help, but not AI purchase decisions [Press release]. Retrieved October 23, 2026, from https://www.gartner.com/en/newsroom/press-releases/2026-05-27-gartner-survey-finds-consumers-want-ai-shopping-help-but-not-ai-purchase-decisions

Kumar, P. (2026). Generative engine optimization at scale: Measuring brand visibility across AI search engines (arXiv:2606.20065) [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2606.20065

Macromill South East Asia. (2026). Perilaku pencarian dengan bantuan AI dalam berbelanja: Peran AI dalam perjalanan belanja konsumen (consumer journey) di Indonesia [AI-assisted search behavior in shopping: The role of AI in the consumer shopping journey in Indonesia] [Survey report]. Retrieved October 23, 2026, from https://www.macromillsea.com/ai-assisted-search-behavior-indonesia

McRoskey, G., Taylor, L., Eppler, B., Barrios, G., Wiener, L., & Lellouche Tordjman, K. (2026, January 2). Consumers trust AI to buy better. Brands need to move quickly. Boston Consulting Group. Retrieved October 23, 2026, from https://web-assets.bcg.com/pdf-src/prod-live/consumers-trust-ai-buy-better-brands-must-adapt.pdf