Even Perfect Generative AI Data Can't Make Decisions? Survey on Purchasing Psychology in the AI Era Reveals Consumers Need 'Human Recommendation' for High-Priced Purchases, Even if They Regret It

Frontier Inc. (Headquarters: Shibuya-ku, Tokyo; President and CEO: Masahiro Takahashi), which operates the business matching service "Ready Crew," conducted a survey targeting men and women aged 20 to 50 on "Consumer Decision-Making and Trust in the AI Era."

The survey polled "Men and women aged 20-50 (n=1,005)" and investigated their awareness of "trust in generative AI and the value of human involvement in decision-making for purchasing and using high-priced goods and services." The results revealed a reality where people seek "human interaction" for final decisions.

【Survey Summary】

1. 【Generative AI's Limitations】 Approximately half responded they "cannot" confidently purchase or use products/services recommended by generative AI.

- The majority responded they "cannot decide (54.0%)" or "less than 50,000 yen (28.2%)" as the amount they can confidently purchase or use products/services recommended by generative AI. Furthermore, about 70% believe that generative AI "does not think it can grasp (65.3%)" complex circumstances not evident in data.

2. 【Human Influence】 Approximately 40% have experience making "high-priced contracts" moved by the enthusiasm of a salesperson.

- About 40% responded they "have (38.6%)" experience signing contracts for high-priced goods or services they did not intend to purchase or use, moved by the salesperson's enthusiasm or attentive proposal. This highlights how "emotional support" that generative AI lacks can change behavior.

3. 【Future Outlook】 The value of "humans" will increase in the generative AI society! Approximately 70% predict that "the value of human involvement will increase more than now."

- In a society where generative AI is widespread, a majority responded they "think the value of trusted human involvement will increase more than now (66.4%)." Regarding satisfaction even in case of regret, "professionals (humans) (74.3%)" were overwhelmingly supported over generative AI (25.7%).

1. 【Generative AI's Limitations】 Approximately half responded they "cannot" confidently purchase or use products/services recommended by generative AI.

When asked, "If you were to make a major purchase or investment that would affect your life (e.g., house, car, high-value investment), would you use generative AI like ChatGPT for information gathering?" approximately half responded "actively use (14.8%)" or "use as a reference (36.8%)."

Although a significant number responded they would "use" it, only about 10% indicated they would "actively use" it, suggesting a cautious stance.

What about the actual purchase or usage decision?

When asked, "What is the boundary amount for which you can confidently decide to purchase or use a product or service recommended by generative AI?" the largest response was "cannot decide based on generative AI's recommendation (54.0%)."

Furthermore, even when a decision can be made, about 30% accounted for "less than 50,000 yen," indicating a trend where trust in generative AI decreases as the price increases.

What kind of concerns do people have about deciding to purchase or use based solely on generative AI's recommendations for high-priced items?

When asked, "What concerns do you have about making high-priced purchases based solely on generative AI's recommendations?" the most common answer was "doubts about the truthfulness and accuracy of information (59.4%)," followed by "information leakage due to inputting personal information into generative AI (28.2%)" and "my own situation and true feelings not being fully understood (27.0%)."

About 60% expressed concerns about the truthfulness and accuracy of information, with concerns about "information leakage due to inputting personal information" and "my own situation not being understood" also ranking high.

To what extent can current generative AI handle "individual preferences" that are difficult to express in data?

When asked, "Do you think current generative AI can sufficiently grasp individual preferences and complex circumstances not evident in data?" only about 30% responded "strongly think so (7.7%)" or "somewhat think so (27.0%)."

These results indicate that the majority feel that current generative AI cannot fully understand the complex circumstances and preferences of humans.

While adept at conditional searches and data comparisons, it is possible that generative AI's limitations are felt in its inability to interpret latent needs and emotional nuances that are not data-driven.

2. 【Human Influence】 Approximately 40% have experience making "high-priced contracts" moved by the enthusiasm of a salesperson.

If the information provided by generative AI were perfect, would human involvement become unnecessary?

When asked, "If you were to make a major purchase or investment that would affect your life (e.g., house, car, high-value investment), even if generative AI presented perfect comparative data, would you want to hear directly from a professional (human)?" approximately 70% responded "strongly want to (23.5%)" or "somewhat want to (42.6%)."

Even with perfect information, many seem to seek direct explanations from human professionals before making a final decision.

For high-priced purchases and investments, the sense of security gained through expert opinions and dialogue likely supports decision-making.

Have there been past shopping experiences where a salesperson's enthusiasm or understanding was the deciding factor?

When asked, "Have you ever had an experience where the final deciding factor for purchasing or using a product or service was the salesperson's enthusiasm or the sense of security that they understood you?" approximately 60% responded "often (12.5%)" or "somewhat often (45.7%)."

It was found that many people have made purchase decisions based on the salesperson's enthusiasm and the feeling of being understood.

These results suggest that in addition to the value of the product or service itself, the salesperson's attitude of being supportive and building a relationship significantly influences the consumer's final decision.

Does this influence extend to deciding on "unexpectedly high-priced purchases?"

When asked, "Have you ever experienced signing a contract for a high-priced product or service that you did not intend to purchase or use, moved by the salesperson's enthusiasm or attentive proposal?" approximately 40% responded "often (5.9%)" or "somewhat often (32.7%)."

Although the majority responded "no," it was found that a certain number of people have experienced purchasing or using unexpectedly high-priced products or services due to a salesperson's proposal.

This indicates that the salesperson's enthusiasm and attentive service have the power to overturn the consumer's initial intentions.

What about the sense of satisfaction if the purchase or use results in regret?

When asked, "If you were to regret purchasing or using a product or service, with whom would you be more satisfied having made the decision to purchase or use: generative AI or a professional (human)?" approximately 70% responded "professionals (humans) (74.3%)."

It was found that even for the same mistake, people feel overwhelmingly more satisfied when the recommendation came from a "professional (human)."

This is likely because the relationship with the salesperson, built through dialogue, creates a sense of satisfaction with the outcome.

3. 【Future Outlook】 The value of "humans" will increase in the generative AI society! Approximately 70% predict that "the value of human involvement will increase more than now."

Even if generative AI handles information gathering and comparison, what roles do you think only humans can fulfill in the final decision-making process for high-priced purchases?

What are the roles that only humans can fulfill in the final decision-making process for high-priced purchases?

- Professional discernment (30s/Female/Shizuoka Prefecture)

- Sensing and understanding emotions from changes in the other person's facial expressions, discerning momentary emotional shifts (tension, relief, etc.) (40s/Female/Hokkaido)

- Calm judgment in complex emotional and environmental situations (50s/Male/Kanagawa Prefecture)

- Difference in proposal ability based on actual experience and specialized knowledge (50s/Female/Nara Prefecture)

Responses highlighting non-verbal communication skills like "sensing emotions from changes in facial expressions" and flexible judgment in complex situations were prominent.

Expectations were also placed on practical knowledge not found in mere numerical data, such as "professional discernment" and "actual experience."

These results suggest that humans are expected to play a role as "companions" who can understand unspoken consumer anxieties and subtle nuances, and propose optimal individual solutions based on their own experience.

Finally, when asked, "In the future society where generative AI becomes widespread, do you think the value of trusted human involvement will increase more than now?" approximately 70% responded "strongly think so (15.5%)" or "somewhat think so (50.9%)."

Even as generative AI evolves and becomes more widespread, many predict that the value of "trusted humans" will increase.

It is thought that precisely because generative AI will handle information processing, the importance of uniquely human roles, such as understanding the other person's emotions and providing responsible encouragement, is being re-recognized.

Conclusion

This survey revealed that while consumers acknowledge the information processing capabilities of generative AI, they are unable to make decisions based solely on its recommendations for high-priced goods and services.

Concerns about the truthfulness of information and consideration for individual circumstances remain strong, and "dialogue and reassurance from professionals (humans)" are sought for final decisions.

In a society where generative AI is widespread, the value of human involvement, which can deeply understand the customer's situation and build trust, will likely increase even further.

【Survey Overview】

Survey Name: "Survey on Consumer Decision-Making and Trust in the AI Era"

Survey Date: May 28, 2026 (Thu)

Survey Method: Internet survey conducted by PRIZMA (https://www.prizma-link.com/press)

Number of Respondents: 1,005 people

Target Audience: Monitors who responded as men and women aged 20-50 at the time of the survey.

Survey Source: Frontier Inc. (https://frontier-gr.jp/)

Monitor Provider: Sakurisa

Business Matching Service "Ready Crew"

Frontier Inc. (https://frontier-gr.jp/), which conducted the survey on "Consumer Decision-Making and Trust in the AI Era," operates the business matching service "Ready Crew," which connects companies.

For high-priced business orders and partner selection where failure is not an option, dedicated business concierges carefully listen to the company's background and complex preferences that AI cannot fully grasp. They support optimal company encounters with "reliable peace of mind" through professional discernment and dialogue that goes beyond simple data searches.

What is Ready Crew?

"Ready Crew" is a free business matching community that handles projects with an annual total transaction budget of 110 billion yen and a cumulative total of 140,000 matching cases. It has been introduced by approximately 70% of listed companies, achieving "optimal matching" that accelerates corporate growth.

Its greatest feature is "Contextual Sales," which uncovers latent issues through dialogue. Through deep listening by business concierges who skillfully draw out challenges even when there are no clear problems, it brings to light the true needs that customers themselves may not have realized.

Ready Crew Usage Examples

https://readycrew.jp/results/

【Company Profile】

Company Name: Frontier Inc.

Headquarters Location: Ebisu Garden Place Tower 17F, 4-20-3 Ebisu, Shibuya-ku, Tokyo

Representative: Masahiro Takahashi, President and CEO

Business Activities: Operation of the business matching agent "Ready Crew"

Founded: November 2009

URL: https://frontier-gr.jp/

【Official SNS】

Facebook: https://www.facebook.com/ReadyCrew

X: https://x.com/readycrew1111

LinkedIn: https://www.linkedin.com/company/78123988/admin/dashboard/

Please note that the press releases and other information posted on this page are current as of the announcement date and are subject to change due to the passage of time or various subsequent events.