[Original Research] Will Your Brand Be 'Recalled' or 'Ignored' by AI?

In the field of SEO and brand management, where professionals are considering how to adapt to AI Overview (AIO), the following sentiments are often heard regarding their company's exposure to AI:

"If it's a non-branded search, it's acceptable if our site doesn't appear." "Isn't AI Overview just for providing general, dictionary-like information?" "As long as users come to our site through branded searches, we don't need to be recognized by AI."

EXIDEA Inc. (Headquarters: Chuo-ku, Tokyo; CEO: Takuma Ogawa) has investigated the extent to which Google Search's AIO generates responses containing brand names for searches where users do not specify a brand. The study covered 17,693 queries and approximately 30,000 brands.

This release shares four key facts that will be useful for SEO and brand managers as they consider their brand's recognition in the AIO era.

Key Findings of This Study

Whether AIO is displayed is more strongly influenced by 'how the question is asked' than by the industry.

AIO tends to be displayed more easily for searches seeking a clear answer, such as 'What is XX?'.

For AIO strategy, it is crucial to prioritize based on the 'query type' rather than treating all keywords uniformly.

Survey Overview

Item Details Survey Target Japanese queries on Google Search that displayed AI Overview and did not include brand, company, or product names (non-branded queries). Number of Analyzed Queries 17,693 Number of Analyzed Blocks 288,177 (paragraph units within AIO responses) Number of Detected Brands 29,950 brands (total of 644,678 mentions) Objective Variable 5-axis classification by LLM (Search Intent / Query Type / Industry / Brand Presence / YMYL Judgment) Brand Extraction Automatic extraction of brand, company, product, and service names from AIO response text by LLM. Statistical Test Chi-squared test p≈0 for all industry, search intent, and query type categories. Kruskal-Wallis test for unique brand count difference H=922.2, p=8.58e-189. Data Acquisition Period February - March 2026 Survey Conductor EXIDEA Inc.

① Even in Non-Branded Searches, Being an AI Comparison Candidate Becomes a Competitive Axis

Even when users searched without including any brand names, 57.7% of AIO response blocks contained specific brand, product, or service names. An average of 4.5 brands were mentioned per query.

On a per-query basis, only 18.8% of queries had no brands appearing in the AIO, while the remaining 81.2% presented at least one brand. Furthermore, in 51.9% of queries, brand names were included in over 60% of the AI response blocks.

Brand Recall Level Condition Query Count Percentage High Brands in >60% of blocks 9,182 51.9% Medium Brands in 30-60% of blocks 3,509 19.8% Low Brands in 1-30% of blocks 1,684 9.5% None Zero brand appearance 3,318 18.8%

AI is evolving from a tool that 'guides users to a specified brand' to one that 'recommends candidate brands on its own, even if not specified.' Relying solely on branded searches as a traffic source may cause businesses to overlook the new competitive axis of whether they are recalled by AI in non-branded searches.

② Prioritize Checking if You Appear as a Candidate in Comparison Searches

The tendency for brand mentions varies greatly depending on the query type. The most significant was in 'Comparison-type' queries, where users search with the premise of 'choosing from multiple options.'

Query Type Brand Mention Rate Avg. Unique Brands Comparison (e.g., XX vs YY, XX difference) 75.9% 6.6 List 61.0% 5.7 How-to 56.4% 4.1 FAQ 52.0% 3.3 Definition (e.g., What is XX) 39.3% 2.7 Reference: Other 56.4% 4.2

In comparison-type queries, three-quarters of AIO responses included brands, with an average of 6.6 brands presented per response. There is a 36.6 percentage point difference in mention rate between comparison and definition types. Among the five clearly classifiable types, AI shows a tendency to actively name brands for 'choosing' searches.

From another perspective, AIO responses to comparison-type queries are akin to an AI-generated 'catalog.' Whether a brand is included as a candidate during comparison is the dividing line for being recalled by AI. Note that 'Other' is a mixed category for queries not clearly classified elsewhere and is treated as a reference value.

③ The Closer the Search is to Purchase, the More Brands AI Names

Looking at search intent, as the user's action stage moves from 'Information Gathering' to 'Comparison' and 'Purchase/Application,' the appearance of brands in AIO responses tends to increase.

Search Intent Description Brand Mention Rate Avg. Unique Brands Navigational Reaching a specific site 73.0% 4.7 Transactional Purchase/Application 68.3% 5.2 Commercial Comparison 64.9% 5.5 Informational Information Gathering 44.7% 3.0

Compared to Informational (44.7%), Commercial (comparison) is 64.9%, and Transactional (purchase/application) is 68.3%. There was a difference of over 20 percentage points between information gathering and purchase consideration.

For searches closer to purchase, the AI has already listed candidate brands before the user clicks. If your company is not a candidate in the comparison phase, you may be eliminated from the options before even getting site traffic.

④ Identify Search Conditions Where AI is Likely to Mention Brands for Each Industry

By industry, the difference between industries where AI is likely to recall brands and those where it is not reached up to 1.9 times. There is a 33.5 percentage point gap between Travel & Tourism (71.2%) and Education & Learning (37.7%).

Industry Brand Mention Rate Avg. Unique Brands Analyzed Queries Travel & Tourism 71.2% 5.4 195 Lifestyle 66.4% 5.3 1,272 Beauty & Fashion 66.1% 5.3 643 Food & Gourmet 65.5% 5.7 240 Finance & Insurance 63.9% 4.8 3,738 IT & Technology 63.1% 5.0 4,420 Housing & Renovation 57.4% 4.6 653 Marketing 55.3% 4.7 1,848 Health & Medical 47.8% 3.6 650 Real Estate 42.7% 3.2 570 Legal & Professional Services 42.0% 2.9 1,211 HR & Career 39.7% 3.2 936 Education & Learning 37.7% 3.5 592

In industries where product/service comparison is common (Travel, Lifestyle, Beauty, Food), AI tends to list candidates more easily. In industries where specialized consultation or individual service is the norm (Legal, HR, Education), AI is less likely to name specific brands. Note that for Travel & Tourism (n=195) and Food & Gourmet (n=240), the sample sizes are smaller, so the figures should be taken as indicative.

Cross-analysis of industry by search intent and query type also revealed hotspots where brand mentions are concentrated. Please use the heatmaps below to check 'under which search conditions AI is likely to list brands' for your industry.

When we compiled the top brands mentioned most frequently in AIO responses by industry, we found that 'mass-market, high-recognition brands' dominated almost every category. It is believed that AI references top-ranking sites in search results, meaning that securing a top search ranking remains a prerequisite for being recalled in the AIO era.

Industry Top 1-3 Brands (Company names withheld) Finance & Insurance Dominated by 2 major consumer finance companies and 1 major credit card company (all with >2,600 mentions) IT & Technology Dominated by 3 mobile communication/wireless internet service companies Beauty & Fashion 1 bridal jewelry company and 2 overseas high-brand jewelry companies Food & Gourmet Dominated by 3 frozen food delivery service companies HR & Career Dominated by 3 HR SaaS vendors Education & Learning Dominated by 3 programming school/online course service companies

Four Points for Managers to Check Immediately

Based on this study, here are four priorities for considering your brand's recognition in the AIO era:

1. Check if your brand appears as a candidate in non-branded searches. Check if your company appears in AIO responses for non-branded searches like '[category name] recommendation,' '[service] comparison,' or 'XX vs YY.'

2. Prioritize checking comparison/recommendation searches. AI tends to list brands in comparison/recommendation searches. First, you should see how your company and competitors are treated in searches where users are looking for options.

3. Find searches where only competitors appear. Searches where competitors appear in AIO responses but your company does not are areas that should be prioritized for improvement. Look not just at 'whether you appear in AIO,' but 'whether you are included among the candidate brands.'

4. Pursue top search rankings and AI candidate inclusion together. AIO strategy is not a replacement for traditional SEO. It is necessary to achieve top search rankings and then ensure you are picked up as a candidate by AI.

EmmaTools Supports SEO Operations in the AI Search Era

This study has shown that in the AIO era, 'how your brand is recalled by AI in non-branded searches' will be a new dividing line for gaining recognition.

EmmaTools is an SEO writing tool that supports the creation, rewriting, and improvement of SEO articles by leveraging search result data like that from this study. We assist with the necessary measures for SEO in the AI era, such as designing content structures that are easily cited by AI, organizing competitive positioning, and prioritizing keywords.

Characteristics of the Research Methodology

This study is characterized by its quantitative analysis of brand appearance trends through large-scale parsing of actual Japanese response blocks generated by Google AI Overview, rather than relying on the subjective views of SEO managers or case studies. It allows for an objective understanding of 'which industries and which searches AI currently names brands for.'

Differentiation Points Details Primary Data Direct analysis of 288,177 Japanese response blocks generated by Google AI Overview, not from general SEO tool APIs. Query Classification Automatic 5-axis classification of each search query using LLM (Search Intent / Query Type / Industry / Brand Presence / YMYL Judgment). Brand Extraction Automatic extraction of brand, company, product, and service names from AIO response text by LLM. Detected 29,950 brands and 644,678 total mentions. Definition of 'Non-branded Query' Queries that do not contain specific company, brand, or product names (e.g., 'credit card recommendation,' 'water server comparison'). Determined by LLM. Statistical Significance Verification Chi-squared test p≈0 for all industry, search intent, and query type categories. Kruskal-Wallis test for unique brand count difference H=922.2, p=8.58e-189. Handling of Reference Category 'Other' in query types is a mixed category and excluded from rankings, treated as a reference. Japan Market Focus Targets only Japanese search results, not global data centered on the English-speaking world.

Note This study is based on a snapshot of Google search results (acquired Feb-Mar 2026), and AI responses may change over time. The extraction of brand/company/product/service names is by automated LLM classification and may contain some errors or omissions. Industry and search intent classifications are also automated by LLM and have certain accuracy limitations. Some cells in the industry-specific and cross-analysis (e.g., Travel & Tourism x Navigational) have small sample sizes (n<30) and should be viewed as indicative. This analysis aggregates brand appearances within AIO blocks.