June 5, 2026

To all the press

GMO TECH, Inc.

GMO TECH deciphers the "grammar of AI recommendations" from 4 domestic AI searches.
How does the generating AI determine what's "perfect for you"?
A research report that will serve as a new benchmark for businesses has been released.

　GMO TECH, Inc. (President and CEO: Ryoichi Suzuki, hereinafter referred to as GMO TECH), a member of the GMO Internet Group, has announced ChatGPT and Google AI Mode.^{(※ 1)}We have published a research report summarizing the trends in how AI-generated product and service suggestions are suggested to users, based on an analysis of 41,264 comparative and evaluation responses in Japanese.
　This study revealed that the generating AI doesn't just suggest "one" answer, but rather offers an average of 4.15 options based on different criteria. This represents a significant change, as it saves users the trouble of comparing vast amounts of information themselves, making it easier for them to make choices that are a better fit for them.

(*1) Google AI Mode: AI summarization and generated response function in Google Search

[Background of the survey]
　The proliferation of generative AI has dramatically improved the way people research information. Previously, users had to manually open and compare numerous websites in search results. However, nowadays, AI often summarizes and compares information on their behalf, allowing users to directly refer to the answers.
　With these changes, for companies, not only "search ranking" but also "under what conditions and contexts their products are introduced or recommended within AI responses" is becoming increasingly important as a new visibility indicator.
　GMO TECH has inductively analyzed and systematized the "structure" of how AI determines "the perfect fit for you" in this new form of information gathering, using actual data.

[Survey Summary]
■ The AI ​​doesn't recommend just "one" option, but suggests choices tailored to the user based on an average of 4.15 criteria.
　AI provides an average of 4.15 options in response to a single question, such as "If you're looking for XX, choose A; if you prioritize YY, choose B." This makes it easy for users to find the option best suited to their situation. For businesses, it becomes crucial to look for opportunities to fit into multiple conditional branches (one row in the recommendation list).

■The main recommendation pattern is "Who is it for?" x "What is it strong at?"
　The main approaches used by AI in its suggestions were "purpose-specific" (37.0%), which focuses on who the information is for, and "function-specific" (34.8%), which focuses on its strengths. Furthermore, a combination of three axes—"beginner/unsure × purpose-specific × function-specific"—was observed in 5.0% of the total, indicating that AI tends to intelligently select and structure products that best suit the user's level and purpose.

■ Rather than "It's a major company, so it's safe," there are more products that cater to the attributes of the user, such as "For people like XX."
　In only 7.1% of AI recommendations were based on traditional appeals to authority, such as "major company," "long-established," or "well-known." Instead, suggestions tailored to the user's attributes and usage scenario (23.8%), such as "suitable for beginners," "classic," or "for people who...", were approximately 3.4 times more frequent. This trend was particularly strong in categories where individual preferences vary, such as cosmetics, hair care, and fashion, highlighting the importance of articulating "who a product is suitable for," regardless of company size.

■ Differences in response structure between ChatGPT and Google AI Mode
　Differences were also observed in the response structure across platforms. Google AI Mode tended to use multiple subheadings and comparison tables even for short questions, offering broad and comprehensive suggestions with an average of 4.58 options. On the other hand, ChatGPT's response structure varied depending on the specificity of the question, and it typically offered suggestions with an average of 3.83 options. Therefore, when companies design content, it is effective to create a structure that targets both platforms by placing a clear conclusion at the beginning and then developing the main text by category and purpose.

■The AI's suggestion criteria also change depending on the genre you are searching for.
　After analyzing 35 genres, we found that the points that AI prioritizes (recommendation patterns) differ from industry to industry.

• Financial Services: Economic Zones, Point Rewards, Risks
• Cosmetics: Skin concerns, user demographics, appearance
• Communication services: Data volume and economic sphere
• SaaS^{(※ 2)}Pricing structure, contract terms, and number of users

(*2) SaaS (Software as a Service): A software delivery model used via the internet.

[Report Summary]

■ Survey outline

• Survey subjects: ChatGPT, Japanese AI responses from Google AI Mode
• Target queries: Questions that include the intention to compare and consider, such as "recommendations," "comparison," "price," "reputation," and "disadvantages."
• Number of analyzes: 41,264
• Number of genres analyzed: 35 genres
• Analysis items: 34 types of recommendation criteria (recommendation axes)
• Data source: Ahrefs Brand Radar API^{(※ 3)}
• Analysis date: May 2026

(*3) API (Application Programming Interface): A mechanism for connecting different systems.

[Future prospects]
　In an era where users use AI to efficiently find products that suit them, companies need to go beyond simply aiming for top search results and instead focus on comprehensive information design that considers "how to have AI explain and suggest products that are suitable for whom."^{(※ 4)}This is required.
　GMO TECH will continue its research and analysis on AI search to support improved consumer convenience and optimal information dissemination for businesses.

(*4) LLMO (LLM Optimization): Optimization measures to increase the likelihood that a company's web content will be referenced in large-scale language models such as ChatGPT.

[About GMO TECH]

　GMO TECH, Inc. is a technology company that leverages AI and SaaS to support businesses in achieving sales growth and improving operational efficiency.
　In addition to developing customer acquisition support businesses in the SEO, MEO, and AI marketing fields, including the store customer acquisition DX SaaS "MEO Dash! byGMO," we also provide operational efficiency support through the real estate DX SaaS "GMO Rental DX," and operate affiliate advertising, internet media, and payment services businesses.
　Leveraging our unique AI technology and in-house developed products, we are promoting the creation of systems that enable AI to continuously deliver value in business operations, supporting companies' digital transformation and sustainable growth.

Above

[Press inquiries]
GMO TECH Holdings, Inc.
　Public Relations Department Kobayashi
　TEL:03-5489-6370 FAX:03-5489-6371
　E-mail: hd-press@hd.gmotech.jp

GMO Internet Group, Inc.
　Group Public Relations Department PR Team Kurata
　TEL：03-5456-2695
　Contact Us:https://group.gmo/contact/press-inquiries/

[Contact information regarding services]
GMO TECH, Inc.
　Product Marketing Division, Sales Department
　TEL: 03-4405-0939 (direct) FAX: 03-5489-6371
　E-mail: pm-marketing@gmotech.jp

[GMO TECH Inc.] (URL:https://gmotech.jp/）

[GMO TECH Holdings, Inc.] (URL:https://hd.gmotech.jp/）

[GMO Internet Group, Inc.] (URL:https://group.gmo/）

Copyright (C) 2026 GMO TECH, Inc. All Rights reserved.

Return to news list