Naver's AI Shopping Agent and Agentic Commerce: AEO Strategy for Korean E-commerce Brands in 2026

In February 2026, Naver launched a Shopping AI Agent beta inside the Naver Plus Store app^[1]. A user types "sofa for a home with a dog," and the agent filters for scratch-resistant materials, summarizes reviews, and guides the purchase through to checkout, all within the same conversation. Since then the agent has gone a step further: it now opens conversations proactively, rather than waiting to be addressed^[2]. This is agentic commerce in practice, and Korea is one of the first markets where it operates at scale. Below is the structure of the shift, Naver's current roadmap, and the concrete AEO steps Korean e-commerce brands need to take to stay visible when an AI is doing the shopping.

Core Concepts: 30-Second Definitions

Agentic commerce is a transaction model in which an AI agent interprets consumer intent and autonomously handles the full purchase cycle (product discovery, comparison, and payment) without the user directing each step.

An AI shopping agent is a conversational AI system that combines a user's natural-language input with their shopping context (history, preferences, budget) to filter a candidate product set, ask clarifying questions, and support a purchase decision.

AEO (Answer Engine Optimization) is the practice of structuring content and data so that an AI agent cites a brand or product when answering a relevant query. In an e-commerce context, AEO is effectively synonymous with "agent-selection optimization."

The Three-Stage Agentic Commerce Flow

Where Naver's AI Shopping Agent Stands Now

Naver has been building toward agentic commerce in stages. The foundation was AiTEMS, its in-house AI recommendation engine, developed in 2017. In March 2025, the company released a dedicated Naver Plus Store app built around AI-driven recommendations^[4]. The conversational Shopping AI Agent arrived in beta in February 2026^[1]: a user describes what they need (or the conditions they care about) and the agent surfaces a discovery guide, walks through product comparisons, and supports additional browsing.

The agent has since moved past passive response. It now initiates conversations with users rather than waiting to be prompted^[2]. The numbers reflect the impact: Naver Plus Store MAU reached 7.77 million in March 2026, an all-time high^[3].

Agent N, announced in November 2025, is the broader strategic frame. It deploys agentic AI across Naver's entire platform (search, reservations, finance), not just shopping. The Shopping AI Agent is its first proof of concept; full rollout across all commerce categories is planned within 2026.

Viewed across the Korean e-commerce market, the rise of the shopping agent is both an opportunity and a structural threat^[5]. A brand that doesn't make it onto the agent's shortlist loses the transaction, irrespective of how much it spends on ads. Gartner forecasts that traditional search engine volume will drop 25% by 2026 (Gartner, 2024)^[6]. For brands whose revenue models depend on search-driven traffic, this is not a background trend; it is a shift in where transactions originate.

Comparing Agentic Commerce AEO Solutions

A set of Korean solutions has emerged to measure and improve brand visibility inside AI shopping agents. The table below summarizes the main options currently available. Prices reflect public listings and are subject to change.

Agentic Commerce AEO: Four Steps for Brands and Sellers

Getting selected by an AI shopping agent requires "data the AI can read and cite," not "ranking first in search results." Four steps form the starting point.

Step 1: Structure Your Product Data

AI agents filter recommendation candidates using product name, category, and attribute fields. Filling in every attribute (material, intended use, target age, compatible specifications) on the Naver Shopping platform is the baseline requirement. On a brand's own site, adding JSON-LD Product schema achieves dual coverage: it lets ChatGPT, Perplexity, and other AI engines pull structured data from the same page, independent of Naver.

A common mistake is polishing the product title while leaving attribute fields blank. AI agents draw on structured attribute values first, before free-text descriptions.

Step 2: Build Content That Answers Conversational Queries

A shopping agent has to respond to situation-driven questions: "sofa for a home with a dog," not "sofa." A product detail page that includes FAQ-style blocks ("When is this the right choice?", "How does it differ from the alternative?") gives the agent extractable content to cite in its answer. Information density on the product page itself matters more than blog posts or series content elsewhere.

Poor example: "Premium sofa crafted from luxury materials": no extractable structure for filtering.
Good example: "Surface material: microfiber (scratch-resistant finish) / Suitable for: homes with pets": maps directly to the agent's filtering conditions.

Step 3: Measure Visibility Across AI Engines

Optimizing for Naver's agent alone is insufficient. As ChatGPT's shopping feed and Perplexity Shopping expand, the same product may be cited differently across platforms, and those gaps need to be tracked. BOIDA (Designovel) tracks visibility across six major AI engines (ChatGPT, Claude, Gemini, and Perplexity among them) simultaneously. Its external validation includes a published paper at ACM CHI 2026 and NVIDIA Inception membership. Tools that track multiple engines in parallel let brands identify engine-specific exposure gaps and prioritize where to act.

Step 4: Refresh Content Based on Citation Patterns

Check which content the AI agent cites (and in what format) on a quarterly basis, then strengthen product information to match the question types and structures that appear most often in answers. Partnering with Korean GEO/AEO specialists such as Next-T, Ascent AI, or Across for regular content-structure audits is another practical option.

What to Do Now

The competitive arena for e-commerce brands has shifted from "rank first in search" to "appear on the AI agent's recommendation list." Naver Plus Store's Shopping AI Agent is the most advanced domestic example of that transition. Three actions cover the ground: structure product data so AI systems can read it, build conversational-query content directly onto product pages, and track visibility across AI platforms regularly. These three are the baseline requirements for competing in agentic commerce.