HubSpot Japan Announces 2026 Customer Service Awareness and Actual Condition Survey

HubSpot Japan Co., Ltd. (Headquarters: Chiyoda-ku, Tokyo; Japan Representative: Yuya Isa; hereafter 'HubSpot'), which provides the AI and CRM-powered Agentic Customer Platform, has conducted the '2026 Japan Customer Service Awareness and Actual Condition Survey' in collaboration with Macromill, Inc. This survey aimed to understand changes in customer service operations, generative AI utilization, and customer inquiry behavior, targeting 618 customer service agents (hereafter 'CS agents') and 206 individuals with experience using corporate customer service (hereafter 'users') in Japan, totaling 824 respondents.

The survey revealed that 88.3% of users 'always' or 'usually' attempt to resolve issues themselves before contacting a company's customer service. Additionally, while 69.9% of users expressed a willingness to use AI-based customer service, top responses for preferring human assistance included situations requiring flexible responses tailored to their circumstances and complex technical troubleshooting. For companies providing customer service, it is crucial to not only create an environment where customers can easily find accurate information before making contact but also to design a customer experience that smoothly transitions between AI and human support as needed.

Survey Highlights

Approximately 90% of users attempt self-service before contacting support: 88.3% of users reported 'always' or 'usually' attempting self-service before contacting customer service. The methods used for self-service are diverse: search engines (69.3%), company FAQ/help pages (56.3%), generative AI (41.2%), and company chatbots (41.2%).

Positive attitude towards AI coexists with expectations for human support: 69.9% of users expressed willingness to use AI-based customer service. However, the top scenarios where human support was preferred were 'when flexible responses tailored to my situation are needed' (47.1%) and 'when resolving complex technical issues' (45.6%).

Higher generative AI adoption among CS agents with organizational support: The overall generative AI adoption rate among CS agents is 35.0%. Among those with organizational support for generative AI use, 67.1% utilize it, compared to only 12.8% in organizations without such support. Furthermore, 56.2% of agents with organizational support view the future of the CS profession positively, compared to 28.0% of those without support.

Survey Target: 618 customer service agents (309 B2B, 309 B2C) and 206 buyers with experience contacting corporate customer service (103 B2B, 103 B2C), all aged 20-69 from the Macromill monitor panel.

Commenting on the survey results, Yuya Isa, Country Manager of HubSpot Japan, stated: 'This survey found that many users attempt to resolve issues themselves using search engines, company FAQs, generative AI, and chatbots before contacting support. This indicates that the customer service experience begins the moment a customer starts searching for information. We believe that creating an environment where customers can quickly find accurate information, and then appropriately connecting them to human support that addresses individual circumstances and emotions, is a key factor determining the customer experience. At the same time, while users are open to AI support, they seek human assistance for flexible, personalized responses and complex technical problems. This aligns with the view among CS agents that AI should handle standard FAQ responses and data aggregation, while humans should handle empathetic interactions. Therefore, it is important not just to introduce AI support, but to design a system that can smoothly transition to human support based on the customer's situation and preferences. Furthermore, we see a trend where organizational support—such as guidelines, training, and sharing success stories—leads to higher generative AI adoption rates and a more positive outlook on the future of the CS profession. Through our Agentic Customer Platform, which combines AI and CRM, HubSpot will continue to support CS organizations in designing customer interactions that leverage the strengths of both AI and humans, from pre-inquiry self-service to utilizing interaction history and human follow-up.'

Survey Results Summary

1. 88.3% of users 'always' or 'usually' attempt self-service before contacting support

When asked how often they try to solve problems themselves before contacting customer service, 48.5% said 'always' and 39.8% said 'usually,' totaling 88.3%. Including those who 'sometimes' attempt self-service (8.3%) brings the total to 96.6%.

Among users who attempt self-service (n=199), the most common methods were 'search engines (e.g., Google)' (69.3%), followed by 'company FAQ/help pages' (56.3%), 'company chatbots' (41.2%), 'generative AI (e.g., ChatGPT, Gemini)' (41.2%), and 'consulting acquaintances/colleagues' (40.7%). Notably, B2B users showed a higher rate of using generative AI for self-service (56.6%) compared to B2C users (26.0%).

These results suggest that customer service needs to be designed not only from the point of contact but from the moment customers begin searching for information. Companies are required to build an information environment—including FAQs, help pages, knowledge bases, and AI-powered initial responses—that allows customers to find accurate official information before making contact.

2. 69.9% of users are willing to use corporate AI support, while scenarios demanding human support are also clear

Regarding AI-based customer service, 69.9% of users expressed willingness to use it. This breaks down to 28.2% who 'actively want to use it' and 41.7% who 'would use it depending on the situation or content.' Conversely, 18.0% preferred human support ('prefer human support if possible' or 'must be human support'). B2B users showed a higher willingness to use AI (74.8%) compared to B2C users (65.0%).

The top scenarios where human support was preferred were 'when flexible responses tailored to my situation are needed' (47.1%) and 'when resolving complex technical issues' (45.6%). Only 16.0% wanted human support in all situations, clearly indicating that while users are open to AI, they seek human assistance based on the context.