Skip to main content
RanketAI
Back to List
geo·Author: RanketAI Editorial Team·Updated: 2026-07-11

ChatGPT Ads Land in Korea — What Advertising Can and Cannot Buy in AI Answers

OpenAI now runs ChatGPT ads in Japan and South Korea — its first non-English markets. We break down the ad format, plan-tier exposure, and targeting, and why the answer body — the spot ads cannot buy — is where Korean brands should invest first.

AI-assisted draft · Editorially reviewed

This blog content may use AI tools for drafting and structuring, and is published after editorial review by the RanketAI Editorial Team.

Key takeaway (as of 2026-07-11): OpenAI has started running ChatGPT ads in Japan and South Korea — its first Asian markets, and the first where the primary user language is not English. Korea is an obvious inventory bet: the ChatGPT app reached 23.45 million monthly active users in April 2026 (WiseApp/Retail), roughly 46% of smartphone users. But what ads can buy is narrow: a clearly labeled card below the response, shown only to Free and Go tier users. The spot where AI mentions and recommends brands inside the answer body is not for sale — and for paid-tier users, that body is effectively the only exposure path. The practical meaning of "the ad era reached Korea" is paradoxical: it just became more urgent to measure the organic visibility that ads cannot buy.

What happened

ChatGPT advertising left the English-speaking world for the first time — and its first stop is Japan and South Korea.

According to ppc.land, OpenAI VP of Monetization Benji Shomair announced on LinkedIn on June 22, 2026 that ads were live in both countries.

"Now running in Japan and South Korea." — Benji Shomair (OpenAI VP of Monetization), via ppc.land

On July 9, OpenAI followed up with an official advertiser email. As Search Engine Land reported, the email paired the Japan/Korea expansion with new Ads Manager features — an Overview tab for account health and performance metrics, suggested ad drafts that prefill proposals from a website's existing images, headlines, and descriptions (notably, it does not generate new copy with AI), custom audiences supporting uploads of 25,000+ users, and a redesigned ad card with larger visuals.

The two markets matter for what they represent. ChatGPT ads launched in the US in February 2026 and reached the UK in June; this is the first move into Asia — and the first into markets where the primary user language is not English. We covered the global split between ChatGPT turning ads on and Perplexity turning them off in a separate analysis — this news means that split now operates on top of Korean-language answers.

Why Korea is the first non-English market

Korea has one of the highest ChatGPT penetration rates anywhere — the inventory case is written in the numbers.

According to a WiseApp/Retail survey reported by ZDNet Korea, Korea's ChatGPT app reached 23.45 million monthly active users in April 2026 — about 46% of the survey's 51.22 million smartphone-user sample. In the same survey, Gemini reached 8.45 million users and Claude 2.41 million. Nearly one in two Korean smartphone users opens ChatGPT at least once a month.

There are signals from the ad-platform side too. Per ppc.land, Criteo said more than 1,000 brands were running ChatGPT ads through its API, with retail conversion rates approaching 2x traditional search — though we note this is an ad platform's own figure without independent verification.

The interpretation is straightforward: the user base long ago passed the size where an ad market becomes viable, and OpenAI has now opened the checkout. For Korean brands the question is no longer "does AI answer exposure concern us?" but "which exposure path do we invest in first?"

The ad format facts — what money can buy

What ads buy is a clearly labeled bottom card, with tight constraints on audience, targeting, and length.

Item Details (per initial announcements)
Placement Labeled ad card at the bottom of the response — advertiser name, favicon, headline, description, URL, image
Recommended length Headlines ~16 characters, descriptions ~32 characters — far more compact than search or social formats
Audience Free and Go tiers only — Plus, Pro, and Enterprise subscribers see no ads
Targeting Conversation-topic context hints, not behavioral profiles — no individual conversation data shared with advertisers
Effect on the answer body None — per OpenAI's official guidance, ads are separate from answer content

The two rows that matter most in practice are the audience and the last one. Ads never reach paid-tier users, and they never touch the answer body. No ad budget, however large, can make the AI say "for this category, consider this brand" inside the answer itself.

The spot ads cannot buy — the answer body

The more the ad era matures, the clearer the value of the one spot that is not for sale.

From a brand's perspective, ChatGPT now has two exposure paths.

  1. The bottom ad card — bought with money. Visible only to Free/Go users, labeled as an ad, and gone the moment you stop spending.
  2. Mentions and recommendations inside the answer body — not for sale. Visible to users on every plan, produced by what the AI has learned and retrieved. It is an asset built by content and trust signals, and it does not disappear when a campaign ends.

For paid-tier users (Plus, Pro, Enterprise), the second path is effectively the only one. If you assume decision-makers who use AI for work skew toward paid plans, then the users ads cannot reach are exactly the users for whom the answer body is your entire brand presence. Why buying your way into that body through gray-market mentions backfires is covered in our paid-mention risk analysis — the legitimate route remains content structured to be cited.

Korea also is not a one-engine market. The same WiseApp/Retail survey counts 8.45 million Gemini users, and Naver keeps expanding its own AI Briefing in search results (see our analysis of Naver AI Briefing's citation economy). Ads opened on one engine; organic visibility is a question across several.

What Korean brands should do now

Whether or not you run ads, the order is the same — measure your organic baseline first, then fill the gaps with content.

  1. Measure your current standing with Korean-language questions. Ask the questions your customers actually ask across ChatGPT, Perplexity, Gemini, and other answer engines, and check whether your brand appears in the answer body — and how often versus competitors. AI visibility tools such as RanketAI focus on measuring exactly this in-answer brand exposure and share versus competitors. Even if you run ads, your baseline is how you appear to the paid-tier users ads cannot reach.
  2. Build content that gets cited. Paragraphs that answer the question directly, verifiable statistics with sources, fresh dates, crawler accessibility — shape your Korean-language content into the form AI systems pick as evidence. We covered the specifics in Building a ChatGPT Answer Block It Will Quote.
  3. Judge ads and organic exposure separately. Ad metrics (clicks on the bottom card) and organic mentions (in-body exposure) are different assets. If you test ads, don't blend the two dashboards — allocate budget based on what remains after the campaign stops.

Start with ads before measuring, and you lose the ability to tell ad-driven traffic from trust built by in-body mentions. The order has to be the reverse — establish the baseline by measurement, then layer ads on top.

Frequently asked questions (FAQ)

Q1. Ads or organic visibility — which should we start with?▾

Measurement first. Ads take effect and vanish the moment you toggle them, but in-answer brand mentions are an accumulated content asset — without a baseline you cannot separate the two. We recommend confirming where your brand appears for Korean-language questions first, then layering ad tests on top.

Q2. What should smaller brands with tight budgets do?▾

Start with the spot that doesn't require ad spend. In-body mentions are determined by content structure (direct-answer paragraphs, statistics, sources, freshness), not budget, and they don't disappear when spending stops. Checking your current state with a free diagnostic tool and reinforcing the weakest question types first is the budget-efficient order.

Q3. Will running ChatGPT ads make our brand appear more in answer bodies?▾

No. Per OpenAI's official guidance, ads appear only in the labeled card below the response and do not influence answer content. In-body mentions are determined by content signals the AI has learned and retrieved, independent of ad spend.

Q4. How is this different from Naver search ads?▾

Different surface, different targeting. Naver search ads are keyword-bid placements above search results; ChatGPT ads are context-hint placements below a conversational answer. And since Naver is expanding its own AI Briefing, Korean brands now manage four distinct surfaces: paid and organic on each ecosystem.

Q5. How do we reach paid-tier (Plus and above) users?▾

Not with ads — Plus, Pro, and Enterprise users see none. The only path to them is mention and citation inside the answer body. The higher your business-user share (typical for B2B), the higher the priority of organic visibility measurement and content reinforcement.

Q6. Should Korean-language answers be measured separately from English?▾

Yes. This expansion made ChatGPT run ads in markets where the primary language is not English, and answer bodies likewise cite different sources and mention different brands depending on the question language. Measure with the Korean-language question set your customers actually use. See our Korean multi-LLM measurement guide.

Update policy

  • Content as of: 2026-07-11 (KST)
  • Update cadence: monthly
  • Next scheduled review: 2026-08-11

Execution Summary

ItemPractical guideline
Core topicChatGPT Ads Land in Korea — What Advertising Can and Cannot Buy in AI Answers
Best fitPrioritize for geo workflows
Primary actionStandardize an input contract (objective, audience, sources, output format)
Risk checkValidate unsupported claims, policy violations, and format compliance
Next stepStore failures as reusable patterns to reduce repeat issues

Data Basis

  • The Japan/South Korea rollout is based on ppc.land's June 2026 report (quoting a LinkedIn announcement by OpenAI VP of Monetization Benji Shomair) and Search Engine Land's July 9, 2026 report on OpenAI's advertiser email. The first-Asian-market / first-non-English-market framing, Free/Go-only exposure, bottom ad-card format with roughly 16-character headlines and 32-character descriptions, and context-hint targeting come from those reports.
  • The July 9, 2026 advertiser email features (Overview tab, suggested ad drafts built from existing website metadata, custom audiences supporting 25,000+ user uploads, and a redesigned static ad card) are based on Search Engine Land's report.
  • Korea's ChatGPT app user base (23.45 million monthly active users in April 2026, about 46% of the smartphone-user sample) is based on a WiseApp/Retail survey reported by ZDNet Korea. Criteo's claim that retail conversion rates approach 2x traditional search is an ad-platform figure without independent verification, and is flagged as such in the body.

Key Claims and Sources

This section maps key claims to their supporting sources one by one for fast verification. Review each claim together with its original reference link below.

External References

The links below are original sources directly used for the claims and numbers in this post. Checking source context reduces interpretation gaps and speeds up re-validation.

Is your site visible in AI search?

See for free how ChatGPT, Perplexity, and Gemini describe your brand.

Start Free Diagnosis →

Related Posts

These related posts are selected to help validate the same decision criteria in different contexts. Read them in order below to broaden comparison perspectives.

Best Lists Rule AI Citations — But Self-Promotion Backfires Unless the Fit Is Right

AI answers cite "best of" listicles more than any format — 43.8% of ChatGPT's cited sources. Yet self-promotion backfires: 43% of answers citing such a page never name the brand. Two Ahrefs experiments show when it works and what to measure.

2026-07-06

One Follow-Up Question Wipes 62% of AI Brand Picks — Measure Persistent Visibility

AI recommendations are unstable: add one buyer detail like "for a small team" and 62% of the brands in the first answer vanish, per Clovion. With SparkToro and ChatGPT's 92.4% referral data, here's why one-shot visibility checks mislead and what to measure.

2026-07-08

Buying AI Citations: The Paid Brand Mention Gray Market and the 2026 Spam Crackdown

A gray market sells brand mentions to game AI citations at 10–15× a backlink. We cover the tactics, the FTC disclosure gap, and Google's June 2026 spam update on AI manipulation — and why measuring earned visibility is now risk management.

2026-06-29

AI Visibility Isn't an SEO Problem — It's Organizational Alignment

McKinsey found 71% of companies use generative AI yet only 39% see real profit impact. When a brand appears wrong in AI answers, the cause is usually inconsistent internal data — not SEO — that confuses LLMs. Here's why, with a four-step alignment playbook.

2026-06-25

When AI Recommends Competitors but Omits Your Brand — Closing the Visibility Gap (2026)

Ask AI "what are the best tools here" and competitors get named while your brand is left out. This guide uses 2026 data on how often LLMs name brands in problem-led queries and what decides who appears, plus a four-step plan to close the gap.

2026-06-16
Back to List