登錄
選單
登錄
海擇短評 Haize Comment：
行程規劃類的新創公司在疫情前與疫情間已有一大票公司終止服務，隨著生成式AI興起，又有一波新血再現江湖；差別在於，過往的行程規劃公司更偏向單純的行程規劃與管理/共享工具，現在則會加入預訂與內容攻略，也算是一定程度的一站式平台。而生成式AI能帶給這類新創公司的價值，目前有兩種比較極端的看法：一部分認為生成式AI能使行程規劃公司建立重要的護城河，而且這項功能在行前/行中都能使用；另一類則認為這類新創公司陷入困境的現況不會改變，因為變現的難度不變。其實，這兩類觀點有一個最大公約數，就是，他們都認同生成式AI至少能帶給旅遊內容新的轉變。
生成式AI能帶給行程規劃的內容什麼改變？過往，新創公司在收集和組織內容的成本非常高，從另一個角度說，就是昂貴。會使用行程規劃的用戶大體都有出境需求，新創公司既需要提供兼具廣度和深度的全球目的地內容，在時間軸上也必須是最新的內容，特別是餐飲與景點的POI，還不時需要更新，這些維護都需要成本；而生成式AI能夠快速的更新相關內容，新創公司要考慮的只有API的對接成本(目前很低)與處理"一本正經的胡說八道"問題。除此之外，如果生成式AI對應的大語言模型內容已同時覆蓋出發地與目的地的語言，它還能處理兩地間的資訊不對稱問題，讓(出發地的)旅客也能直接取得目的地的內容。這意味著，過往不同語言的跨境旅遊內容依賴KOL與媒體推廣，而現在旅客能通過生成式AI直接取得目的地的POI內容；舉例來說，這也意味著，受當地居民歡迎的餐廳，未來不需要太多推廣也能得到更多國際遊客的造訪。
不過，生成式AI對行程規劃內容上的影響，未必能帶給新創公司良好的變現能力。歸根究柢來說，問題在於，一站式的旅行規劃工具與附加行程規劃工具的旅遊預訂或攻略網站有何不同？公司是否能盈利，仍與往昔海擇資本討論的議題息息相關，比如：所面對的潛在市場規模、商業模式是否可行、團隊執行力是否高效、如何在低頻的旅遊行業生態增長用戶的生命週期，這些都是老問題，且全世界的旅遊公司都得面對。後疫情時代一些旅遊內容的生成與傳播方式發生改變，現在對接API的技術成本也比過往更低、更友好，但這些本質上的問題不會改變。
海擇資本對生成式AI投贊成票，它正在協助旅行者在靈感發想與行程規劃階段，做到很多過去做不到的事，未來還會做到更多。不過，新創公司是否能通過生成式AI，在旅遊領域規劃建立穩定的商業模式，則是另一件事；即便目前很多旅遊預訂領域的巨頭與ChatGPT合作，也談不上貢獻極大的收入。至於，最後真正受惠的公司，也許不是新創公司，不是OpenAI，不是成熟的預訂巨擘，而是同時擁有跨領域資源的Google Map+Gimini+Bard。
-------
Before and during the pandemic, numerous travel planning startups ceased operations. With the rise of generative AI, a new wave of these companies has emerged, offering not just itinerary planning and management/sharing tools, but also booking services and travel guides, effectively creating a one-stop platform. Opinions on the value of generative AI for these startups vary widely. Some believe it provides a significant competitive advantage that can be utilized both before and during travel, while others see no change in the challenges these companies face, particularly in monetization. However, there's a consensus that generative AI can at least bring about fresh changes to travel content.
What changes can generative AI bring to travel planning content? Traditionally, the cost for startups to collect and organize content has been very high. Travel planning users generally have international needs, requiring up-to-date content on global destinations with both breadth and depth, especially for dining and sightseeing points of interest (POIs), which frequently need updates—all costly to maintain. Generative AI can quickly refresh related content, with startups mainly considering the low cost of API integration and managing potential inaccuracies. Additionally, if the AI’s underlying large language model covers both the language of the departure and the destination, it can address information asymmetries, allowing travelers to directly access destination content. This reduces the reliance on influencers and media for promoting cross-border travel content in different languages, potentially making local favorites more accessible to international tourists without extensive promotion.
However, the impact of generative AI on travel planning content may not significantly enhance monetization for startups. The core issue remains: what distinguishes a one-stop travel planning tool from a traditional travel booking or guide website with added itinerary features? Profitability still revolves around long-standing debates familiar to Haize capital, such as potential market size, viability of the business model, team efficiency, and increasing the customer lifecycle in a low-frequency travel industry. These are enduring challenges faced by travel companies worldwide. While post-pandemic changes in content generation and distribution, and lower, more accessible API integration costs are noteworthy, they do not alter these fundamental issues.
Haize Capital supports generative AI, which is already enabling travelers to do things during the inspiration and planning phases that were previously impossible, with more advancements expected. However, whether startups can develop stable business models in the travel sector through generative AI remains to be seen; even as many travel booking giants collaborate with ChatGPT, significant revenue contributions are not yet evident. Ultimately, the real beneficiaries may not be the startups, OpenAI, or the established booking giants, but rather entities like Google Maps+Gimini+Bard that possess cross-domain resources.
標籤 Label： Generative AI Travel planning ChatGPT Google