Naver and Kakao are moving faster to use artificial intelligence to handle everything from information searches to place recommendations and bookings within their own platforms.

Naver is focusing on linking questions that start in its search box to shopping, places and bookings. Kakao is focusing on connecting travel and gathering demand identified in KakaoTalk conversations to KakaoMap and its booking service.

The two companies start from different points, search and conversation, but they share the same goal. That is for AI to do more than answer questions, by identifying user intent, calling internal services and leading to real actions.

Naver links search results to maps and real-time bookings

Naver's AI tab is an expanded version of its integrated search. It replaces the existing "Green Dot" button in the search box and was officially launched on June 26.

If a user enters multiple conditions in a single sentence, such as time, party size and area, AI recommends places by combining place information, visitor reviews and blog posts. Users can stay on the search screen and continue to check maps and view available time slots for real-time reservations.

Naver is also releasing metrics to support its ability to drive actions. During the beta service period, click-through rates for product cards and place cards each topped 20 percent. Users who visited the AI tab at least 11 times clicked products 2.7 times more and places 2 times more than users who visited once.

Since the official launch, the daily average number of queries has risen 7 times from the beta period, and queries per user have increased 1.7 times. For searches with many conditions, such as electronics or household goods, Naver said it cut the time from search to decision-making by up to 60 to 70 percent. Against that backdrop, AI tab users topped 10 million on July 15.

Naver is also broadening touchpoints. It added an AI tab chat window at the bottom of "AI Briefing", used by more than 30 million people a month, so that users who check a summary can move directly to follow-up questions and then bookings or purchases.

Kakao expands place agent beyond restaurants to travel and everyday life

Kakao has also recently upgraded travel features in "Kanana in KakaoMap" and "Kanana in KakaoTalk". A key feature is that it starts in KakaoTalk chat rooms rather than a search box.

When friends or family discuss travel schedules and destinations in a chat room, Kanana identifies trip length, number of people, departure area and preferred activities from the conversation context and suggests destinations. Once a destination is set, it proposes places and travel routes in KakaoMap, and connects accommodations and restaurants listed on KakaoTalk Booking to reservations on the spot. The recommended itinerary can be shared back in the chat room so companions can review it, ask additional questions or take part in reservations.

The scope of this conversation-based recommendation is also expanding quickly. Kakao expanded its place agent in "Kanana in KakaoTalk" beyond restaurant recommendations and reservations to tourist sites, accommodations, exhibitions, performances and cinemas, as well as gas stations, convenience stores and auto repair shops, through an update on June 29.

Even a conversation about weekend plans with a friend can lead Kanana to suggest pop-up exhibitions or nearby day-trip destinations first.

Kakao's strength is that it can tap demand even before users start a search. It is because AI can step in from the moment people discuss in a chat room, "Where should we go on vacation," before entering a destination into a search box.

Different starting points in search and chats, but aimed at bookings and purchases; the battleground is converting into action

Naver and Kakao use different approaches but share the same destination. Naver is building a structure that runs "from search to action", while Kakao is building one that runs "from conversation to action". The key in AI competition is not answer quality but how much it leads users to actual reservations or purchases. That is also why Naver is repeatedly disclosing behavioral indicators such as click-through rates and differences by usage frequency.

The battleground is also widening beyond travel and places. Naver will introduce a real estate listings recommendation agent and a Whale browser-only agent this month, and add a health agent within the year. Kakao also plans to expand links to internal services and cooperation with external partners to broaden Kanana's activity beyond travel and everyday convenience.

The securities industry view is that it remains to be seen whether these plans will lead to actual results.

So-hye Kim (김소혜), an analyst at Hanwha Investment & Securities, said Naver's integrated service assets that run from search to purchases and bookings could be a strength. But she said at least two quarters of verification are needed to check whether it is cannibalising existing advertising demand and to confirm actual conversion rates.

Yu-jin Seon (선유진), an analyst at LS Securities, also said Kakao's agent service must prove optimised targeting, a differentiated user experience and actual contributions to profitability to be recognised as AI monetisation.

Ultimately, the AI booking race between Naver and Kakao depends less on who delivers more natural answers than on how much they convert user intent captured in searches and conversations into actual reservations and purchases. The next battleground is likely to be which side delivers higher action conversion and transaction performance.