Recommended Hotel Survey (Hakone Edition) Conducted - 114 Accommodations Appeared in Responses from 5 Generative AIs, 31.1% of Top 5 Recommendations Concentrated

Stellagent Inc. (Headquarters: Yokohama, Kanagawa Prefecture, CEO: Akihiro Suzuki, hereinafter "the Company"), which develops an agentic commerce foundation supporting purchasing and reservation experiences in the AI agent era, is pleased to announce the results of the "AI Recommended Hotel Survey (Hakone Edition)" to investigate which hotels and inns major generative AIs recommend as accommodations for Hakone travel.

In this survey, we asked five AI services – ChatGPT, Claude, Gemini, Perplexity, and Microsoft Copilot – to recommend hotels and inns in Hakone five times each for five travel scenarios: couples travel, family travel with children, inbound travel, student/youth travel with friends, and travel with parents/extended family. As a result, out of a total of 125 responses and 625 recommendation slots, 621 recommendations were extracted as accommodation names, and 114 accommodations appeared. According to Hakone Town data, there were 454 inns and hotels in the town as of Reiwa 6 (2024), and the 114 accommodations appearing in the AI responses represent approximately 25.1% for comparative reference. Furthermore, the top 5 accommodations appeared a total of 193 times, accounting for 31.1% of the 621 valid recommendations. The most frequently appearing accommodation was Fujiya Hotel with 51 mentions, followed by Hakone Kowakien Ten-yu with 49, Hakone Ashinoko Hanori with 34, Tenseien with 30, and Hakone Yutowa with 29. As AI is becoming an entry point for travelers choosing hotels, it is thought that not only search result and OTA rankings but also which travel scenarios accommodations are recommended for and the reasons for those recommendations will become new points of focus for attracting customers.

Survey Results Summary

5 AIs x 5 Scenarios x 5 Times = Total 125 Responses, 625 Recommendation Slots Obtained

621 recommendations extracted as accommodation names, 114 accommodations appeared

Compared to 454 inns and hotels in Hakone Town as of Reiwa 6 (2024) according to Hakone Town data, the 114 accommodations appearing in AI responses represent approximately 25.1% for comparative reference.

The top 5 accommodations appeared a total of 193 times, accounting for 31.1% of the 621 valid recommendations.

Most recommended were Fujiya Hotel (51 times), Hakone Kowakien Ten-yu (49 times), Hakone Ashinoko Hanori (34 times), Tenseien (30 times), and Hakone Yutowa (29 times).

Top 1 recommendation counts: Gora Kadan (19 times), Hakone Hotel Kowakien (16 times), Fujiya Hotel and Hakone Yutowa (13 times each) were top-ranked.

Hakone Hotel Kowakien was prominent for family travel with children, Gora Kadan for inbound travel, and Fujiya Hotel for travel with parents/extended family.

Perplexity returned responses that included booking sites, groups of inns, and feature pages as recommendation slots for 4 instances, rather than specific accommodation names.

Survey Background: Hotel Selection Shifts from "Search" to "Consulting AI"

Travelers are increasingly consulting AI not only by entering area, dates, number of people, price, and facility conditions on hotel booking sites for comparison, but also by asking questions like, "Where should we stay in Hakone for a couple's trip?" "Which inn is good for a family trip with elementary school children to Hakone?" or "Which ryokan in Hakone is suitable for a couple visiting Japan for the first time?"

In such cases, the facilities displayed in the AI's response become the traveler's initial candidate list. For hotel and inn operators, it is crucial not only for their official websites to be found in search results but also for AI to correctly understand the facility's features and recommend it to the appropriate travelers. To clarify new points of consideration for hotel customer acquisition in the AI agent era, our company conducted a survey of major AIs' recommendations for hotels and inns in the Hakone area.

Item

Details

Survey Name

AI Recommended Hotel Survey (Hakone Edition)

Survey Target

Hakone Hotel and Inn Recommendation Responses from Major Generative AIs

Target AIs

ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot

Survey Scenarios

Couples Travel, Family Travel with Children, Inbound Travel, Student/Youth Travel with Friends, Travel with Parents/Extended Family

Number of Executions

5 AIs x 5 Scenarios x 5 Times = 125 Responses Total

Recommendation Slots

625 Slots

Recommendations Extracted as Accommodation Names

621 Cases

Survey Period

July 1-2, 2026

Survey Method

Inputting identical prompts into each AI's web UI and recording recommended facilities, recommendation order, reasons for recommendation, and presented booking routes.

Suppression of Contextual Influence

Used Temporary mode, temporary chat, or new chat for each AI service. Explicitly stated in prompts not to refer to past conversation content or memory.

Survey Entity

Stellagent Inc.

*This survey is an exploratory study based on AI responses as of the survey date. AI responses may vary depending on the date, model, search settings, conversation history, region, account settings, etc. This survey does not evaluate the superiority or inferiority of each hotel or inn.

Survey Result 1: AI Responses Concentrated on 114 Accommodations, Top 5 Facilities Accounted for 31.1%

When we asked 5 AIs for recommended hotels and inns in Hakone five times each for five travel scenarios, we obtained 625 recommendation slots, of which 621 recommendations were extracted as accommodation names. A total of 114 accommodations appeared, with the top 5 facilities appearing 193 times, accounting for 31.1% of the 621 valid recommendations.

According to Hakone Town data, there were 454 inns and hotels in the town as of Reiwa 6 (2024). The 114 accommodations appearing in the AI responses represent approximately 25.1% for comparative reference. However, since AI responses may include facilities outside of Hakone Town or those requiring location verification, this percentage is a preliminary figure before detailed verification of facility locations.

*The number of inns and hotels in the town is based on "Document 1: Current State of Tourism and Financial Outlook" presented at the "2nd Hakone Town Tourism and Community Development Revenue Study Meeting (held on August 6, Reiwa 7 (2025))."

Rank

Facility Name

Number of Recommendations

Number of Top 1 Recommendations

Main Recommended Scenarios

1

Fujiya Hotel

51

13

Travel with Parents/Extended Family, Family Travel with Children

2

Hakone Kowakien Ten-yu

49

10

Family Travel with Children, Travel with Parents/Extended Family

3

Hakone Ashinoko Hanori

34

9

Couples Travel, Family Travel with Children

4

Tenseien

30

1

Family Travel with Children, Youth Travel with Friends

5

Hakone Yutowa

29

13

Youth Travel with Friends, Couples Travel

Fujiya Hotel was frequently recommended for travel with parents/extended family and family travel with children. Hakone Kowakien Ten-yu consistently appeared for family travel with children and travel with parents/extended family. Additionally, although Hakone Hotel Kowakien ranked 7th in total recommendations, it had the second-highest number of top 1 recommendations with 16, and it was particularly prominent for family travel with children, indicating that facilities strong in specific scenarios differ.

Survey Result 2: Recommended Facilities Vary Significantly by Scenario

Even for the same "recommended hotels and inns in Hakone," the facilities chosen by AI differed depending on the travel scenario.

Scenario

Facilities with High Recommendation Counts

Couples Travel

Tsuki no Yado Sara, Hakone Ashinoko Hanori, Hakone Kowakien Ten-yu, Hakone Yutowa, Fufu Hakone

Family Travel with Children

Hakone Hotel Kowakien, Hakone Kowakien Ten-yu, Tenseien, Hotel Okada, Fujiya Hotel

Inbound Travel

Gora Kadan, Hakone Ginyu, Yama no Chaya, Hakone Kowakien Ten-yu, Hyatt Regency Hakone

Student/Youth Travel with Friends

Hakone Yutowa, Tenseien, Fujiya Hotel, Hakone Ashinoko Hanori, Hakone Hotel

Travel with Parents/Extended Family

Fujiya Hotel, Hakone Kowakien Ten-yu, Kin no Yu Setsugekka, Hakone Suishouen, Hakone Yutowa

AI responses showed a tendency for evaluation criteria such as formality, accessibility, room usability, in-facility amenities, dining style, and hot spring experience to vary within the same area depending on "who you are traveling with." For accommodation facilities, understanding which travel scenarios they are likely to be recommended for is important for information design in the AI era.

Survey Result 3: Recommendation Bias Differs by AI Service

Even with the same prompt, the number of recommended facilities and top-ranked facilities differed by AI service.

AI Service

Number of Recommended Accommodations

Facilities with High Recommendation Counts

ChatGPT

32 Facilities

Hakone Yutowa, Hakone Suishouen, Fujiya Hotel

Claude

35 Facilities

Fujiya Hotel, Hakone Yutowa, Hakone Kowakien Ten-yu

Gemini

32 Facilities

Hakone Kowakien Ten-yu, Fujiya Hotel, Hakone Ashinoko Hanori

Perplexity

51 Facilities

Tenseien, Tsuki no Yado Sara, Fujiya Hotel

Copilot

31 Facilities

Hakone Ginyu, Hakone Ashinoko Hanori, Hakone Suishouen

Perplexity had the highest number of appearing facilities at 51, with recommendations relatively dispersed. On the other hand, Copilot and ChatGPT showed a tendency for recommendations to be concentrated on a more limited set of facilities. Additionally, for 4 slots in Perplexity, the responses included booking sites, groups of inns, and feature pages rather than specific accommodation names.

Stellagent's View: "Structuring Facility Information" and "Visibility by Scenario" are Important for Being Chosen by AI

In an era where AI agents assist travelers in choosing hotels, accommodation facilities are required to take the following actions:

Organize facility information, room information, hot springs, dining, accessibility, child-friendliness, and suitability for group travel or travel with parents in a way that AI can easily understand.

Periodically check which travel scenarios the company is likely to be recommended for, or not recommended for.

Prepare official information that AI can easily reference, and reduce inconsistencies in the notation of facility names, locations, rooms, hot springs, dining, and booking routes.

Connect AI-driven travel consultations to official booking routes, membership programs, and plan comparisons.

Ensure the freshness and structure of public information so that AI does not provide incorrect explanations.

Responding to AI in hotel bookings is not just about introducing FAQ chatbots. Information design to be chosen by AI and route design to connect from AI to official bookings are the next customer acquisition challenges.

Detailed Report and Consultation Desk

The full text of this survey, prompts, aggregation methods, and recommendation trends by model and scenario are published on the public survey report page on our company's website.

Detailed Report Here: https://stellagent.ai/ja/research/hakone-hotel-ai-recommendation-survey-202607

Stellagent offers consultations for hotels, inns, and tourism operators on designing official booking routes for the AI agent era.

1. Consultation on AI Recommendation Status and AI Countermeasures

For hotels and inns in the Hakone area who wish to confirm how their facilities are recommended to which traveler scenarios on AI, how they are explained compared to competitors, and whether their official websites and booking routes are correctly referenced based on the survey results, we provide AI recommendation status diagnosis and improvement consultations.

2. Consultation for Regional AI Recommendation Surveys

In regions other than Hakone, we can survey which hotels, inns, and tourist facilities major AIs recommend, and which facilities are likely to be candidates by travel scenario. We design and conduct AI recommendation surveys for municipalities, DMOs, tourism associations, hotel groups, and business consortia, setting regional names and travel scenarios.

About Stellagent

Stellagent Inc. develops an agentic commerce foundation that supports purchasing and reservation experiences in the AI agent era. For a world where AI agents search, compare, reserve, purchase, and pay for products and services on behalf of users, we provide the technological foundation that allows businesses to make their product information, inventory, membership base, reservation systems, and payment functions available to AI agents.

In addition to the retail and EC domains, we develop agentic commerce foundations, support AI agent development, and conduct surveys and diagnoses of AI agent readiness to enable companies to build new customer touchpoints via AI agents in reservation domains such as railways, airlines, and hotels.

Company Name: Stellagent Inc.

Representative: Akihiro Suzuki

Location: WeWork, 8F Ocean Gate Minatomirai, 3-7-1 Minatomirai, Nishi-ku, Yokohama, Kanagawa Prefecture

Establishment Date: February 22, 2022

HP: https://stellagent.ai/ja

*Product and service names mentioned in this press release are trademarks or registered trademarks of their respective companies.