[AI Search Usage Survey] 72% of Western Users Have Used AI Search: Trust Effect of AI Adopters is 9x Higher Than Non-Users, 50 US Respondents Show New Norms of the LLMO Era

IGNITE Inc. (Location: JAM-STUDIO 912, 9F JAM BASE, Grand Green Osaka North Building, 6-38 Ofuka-cho, Kita-ku, Osaka 530-0011; CEO: Yoshiaki Kosuge) conducted a 'Survey on Attitudes Towards Western AI Search and LLMO Strategies' targeting 50 men and women residing in the Western United States from February 9-16, 2026. The survey revealed three key findings that fundamentally challenge the assumptions of many Japanese companies regarding their AI search strategies for the West: (1) The percentage of AI users has reached 72%, shattering the premise that it is a tool for a niche group of enthusiasts. (2) A sharp divide in perception: 88.9% of early AI adopters report a significant increase in trust from AI recommendations, whereas 71.4% of non-users state it is 'not a basis for trust.' (3) A negative reaction from 54.5% of mixed-use users (who use AI depending on the situation), who perceive LLMO-focused appeals as 'manipulative'.

Survey Result Highlights

This survey clarified the following three points.

Evaluation of LLMO measures differs by up to 9 times depending on AI usage depth

While 72% of respondents are AI users, it was found that evaluations of the same LLMO measures are sharply divided by segment. Among early AI adopters, trust in AI recommendations reached 88.9%, whereas 71.4% of non-users responded that it is 'not a basis for trust.' This creates a 9-fold difference in the trust effect of the same measure, indicating a risk that designing policies based on the average '72% AI user' figure could be counterproductive for some target audiences.

Company-proprietary data significantly improves favorability among AI users, but does not reach non-users

When company-proprietary data was cited in AI responses, 66.7% of early AI adopters and 45.5% of mixed-use users reported a 'significant increase in favorability.' On the other hand, 85.7% of non-users stated it had 'no particular effect.' While preparing proprietary data was confirmed to be a highly effective approach for reaching AI users, it was also clarified that it is not a measure that functions equally for everyone.

Evaluation of LLMO measures is polarized depending on AI usage depth

Regarding the evaluation of companies implementing LLMO measures, completely opposite results were confirmed: 77.8% of early AI adopters evaluated them favorably as 'sincere and cutting-edge companies,' while 54.5% of mixed-use users showed a negative reaction, calling them 'unnatural and manipulative.' This shows that the evaluation reverses completely even among AI users depending on their depth of use, indicating that 'how' LLMO measures are communicated is a crucial factor that influences brand evaluation.

Survey Background

At IGNITE, we are receiving an increasing number of inquiries from Japanese companies considering content marketing for Western markets, with questions like 'How can we get our brand to appear in AI chatbots?' and 'We don't know how to prioritize LLMO measures.' A recurring issue that emerges from these consultations is that 'reactions to the same measure differ completely depending on the target's AI usage level.'

This survey was designed as a comprehensive flow, going beyond the frequency of AI tool usage to include 'how AI recommendations are received,' 'the impact of not being featured in AI Overviews,' 'evaluation of company-proprietary data,' and 'impressions of companies with LLMO measures.' Despite the n=50 scale, it covers all age groups evenly and provides a solid empirical starting point for planning LLMO strategies.

Q1: Frequency of AI search tool usage | 'Only 28% are non-users,' AI usage becomes daily life in the West

Question: 'When gathering information, such as for travel planning or researching business tools, do you use AI tools like ChatGPT or Perplexity before checking Google search results?'