Naver rolled out an AI Tab on June 26, 2026, replacing its green search box with a conversational search interface, part of a broader shift by South Korea's two largest portals toward agent-style search. Asiae reports that answers from Naver Knowledge iN contributors designated as "Naver Mates" were cited about 60 million times in AI briefings over the past month, and MK News reports a 36% one-week return rate for users in the AI Tab beta. Rival Daum, now operated by AXZ, launched its own AI Overview feature on July 1 using the Solar model built by Upstage, which acquired AXZ from Kakao in May. Naver plans to introduce advertising inside AI Briefing next month, a monetization move practitioners in retrieval, ranking, and ad systems should watch closely.
The shift from single-turn chat answers to agent-style, multi-turn search interfaces changes the telemetry available to engineers and product teams: session depth increases, intents become multi-step, and the mix of signals feeding ranking and ad systems shifts accordingly. That matters directly for practitioners responsible for relevance evaluation, session-based modeling, and ad integration, and it is playing out now in South Korea's search market.
What happened
Naver launched its AI Tab on June 26, 2026, replacing its green search box with a conversational interface that aggregates results from search, shopping, places, blogs, cafes, and clips into a single flow, according to MK News. Asiae reports that answers sourced from Naver Knowledge iN contributors designated as "Naver Mates" were cited about 60 million times in AI briefings over the past month, and MK reports a 36% one-week return rate for users during the AI Tab's beta test. Rival portal Daum, now operated by AXZ after Upstage acquired AXZ from Kakao in May, launched its own AI Overview feature on July 1 that uses Upstage's Solar model, presents summaries with supporting evidence, and auto-updates results as source material changes, per Asiae.
Industry context
MK reports Naver's search market share rose from 63.82% (Jan 1-Apr 26) to 66.34% (Apr 27-Jun 17) and surpassed 80% on June 24, underscoring the scale at which any interface change to its core search product plays out. Asiae adds that Upstage's role as Daum's model supplier illustrates a broader pattern: portals without an in-house frontier model can move quickly into agent-style search by integrating a third-party foundation model rather than training one from scratch.
For practitioners
Teams building retrieval, ranking, or ad systems should watch for three shifts as agentic interfaces roll out: per-citation provenance and confidence scoring becoming a production requirement, offline evaluation expanding from single-response CTR to session-level satisfaction and multi-step task completion, and monetization experiments needing new frameworks once ads move inside conversational answers rather than static result pages. MK reports Naver plans to introduce advertising inside AI Briefing starting next month, which will be an early real-world test of ad relevance and click-through inside a conversational format.
What to watch
Whether Naver publishes interfaces or APIs exposing citation confidence and provenance for downstream tooling, whether session-level metrics replace single-CTR targets industry-wide, and how advertisers and users respond once Naver's AI Briefing ads go live next month.
Key Points #
- 1Naver replaced its search box with a conversational AI Tab on June 26, aggregating results across verticals into one agent-style flow.
- 2Rival Daum launched a competing AI Overview on July 1 using Upstage's Solar model, following Upstage's May acquisition of operator AXZ.
- 3Naver plans to add advertising inside AI Briefing next month, testing how ad relevance and monetization work in conversational search.
Scoring Rationale #
Notable product rollout in a major regional search market with measurable engagement signals (citation volume, return rate) and an imminent ad-monetization change, corroborated across three independent Korean business outlets. It matters for practitioners building retrieval, ranking, and ad systems in agent-style search but is a regional portal story rather than a frontier-model event.
Sources #
Public references used for this report. Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
[High-Value Overnight OrdersEasy](/problems/sql/high-value-overnight-orders)
[Delivered International ShipmentsMedium](/problems/sql/delivered-international-shipments)
[On-Time Delivery Rate by CarrierHard](/problems/sql/on-time-delivery-rate-by-carrier)
250 free problems · No credit card
See all Logistics & Shipping problems