Naver is preparing to make the advertiser’s landing page an active source of search-ad creative. Beginning with a test on September 29, its Powerlink system can automatically generate additional AI summary text from existing ad assets, landing pages and business-channel information, adapting the message to the user’s search intent. Advertisers will not write or edit these generated descriptions themselves; Naver’s model will assemble them dynamically and display them alongside the advertiser’s existing creative when the system decides they are relevant.
The timetable comes from Naver’s official September 22 update. The test will run from September 29 through October 6 on a subset of traffic and queries. From October 7, advertisers will be able to enable or disable the feature at ad-group level, and formal deployment is scheduled for October 13 across eligible Powerlink placements in both desktop and mobile integrated search. Naver notes that the schedule can still change for internal operational reasons.
The landing page becomes creative input, not just the destination
Naver calls the feature “AI summary text.” Its purpose is to add a query-relevant summary to the creative the advertiser has already registered, helping users understand the advertisement and potentially move toward a click. The generated text can draw from the advertiser’s ad materials, landing page and business-channel information. Naver says it will apply automated checks intended to prevent inaccurate or inappropriate expressions, while also explicitly stating that it does not guarantee the accuracy or appropriateness of generated text.
That changes the operational role of the landing page. It is no longer only the page a user reaches after clicking an ad. Its language can become upstream input to the message shown before the click. Product descriptions, service details, claims, positioning and other page content therefore have the potential to influence both the post-click experience and the AI-generated pre-click creative.
This closely parallels a shift NetContentSEO has been tracking across AI advertising systems. In our analysis of ChatGPT Ads, OpenAI’s Ads Manager similarly uses landing pages as source material for AI-assisted creative, while landing-page content can also participate in relevance and eligibility decisions. Naver’s implementation is different, but the direction is increasingly familiar: websites are becoming machine-readable campaign infrastructure.
The generated copy is labeled AI—and advertisers cannot edit it
Naver says users will be able to distinguish the generated message from conventional advertiser-authored creative through an “AI ⓘ” label. Tapping the information icon will provide additional explanation. The AI summary will appear in addition to existing ad creative rather than replacing the normal registration process.
Advertisers, however, cannot directly enter or modify the AI summary text. If a generated message does not accurately reflect the business, Naver’s guidance is to keep the underlying advertising information and landing page accurate, or disable the feature for the relevant ad group once that control becomes available. That makes source-data quality more consequential because the correction mechanism is indirect: fix what the model reads or turn the generation feature off, rather than rewriting the generated sentence itself.
The opt-out is also notable because it will not exist during the September 29–October 6 test. Naver says the setting will be unavailable during that period for performance-comparison purposes. From October 7, advertisers can control the feature through the ADVoost Max creative-generation setting at ad-group level and can disable it before the October 13 formal rollout. Naver says opting out will not change existing ad exposure, ranking calculations or billing.
Search intent becomes part of the creative-generation loop
The system is not designed to show the same generated summary every time. Naver says exposure can vary according to the search query and the feature’s contribution to advertising performance, and AI summary text will not necessarily accompany every eligible Powerlink impression. During the initial test, only some traffic and some search terms will receive it.
This moves paid search another step away from the idea of a fixed advertisement matched to a keyword. The platform can interpret the query, inspect the advertiser’s own information and construct an additional message intended to bridge the two. The advertiser supplies source material; the search system decides whether to generate a message and what that message should say.
Google is moving in a related direction with automated Search advertising. NetContentSEO recently examined Google’s effort to expose the query-to-creative-to-landing-page journey in AI Max. Across platforms, the traditional boundaries between query targeting, creative selection and destination-page interpretation are becoming less distinct as AI systems participate in all three stages.
Naver will show advertisers what its AI actually displayed
The reporting component may be as important as the generation itself. From October 13, advertisers will be able to open ADVoost Max Insights and inspect AI summary texts that were shown with their ads during the previous seven days. This gives marketers a retrospective record of the language users actually encountered rather than leaving dynamic generation completely opaque.
That visibility matters because an advertiser cannot manually approve or edit each generated sentence before it appears. A seven-day history creates an auditing surface: teams can review how the model interpreted their products, services and landing pages, identify recurring messages, spot problematic summaries and decide whether the underlying source content needs correction or the feature should be disabled.
It also creates a new kind of creative research. Generated messages can reveal which aspects of a landing page Naver’s system considers useful enough to surface for real queries. Advertisers should be careful not to interpret this as a complete record of Naver’s ranking logic, but the report can still expose the practical language being produced at the intersection of search intent and site content.
Clicks on AI text remain normal paid clicks
The addition does not create a separate auction or billing model. Naver says ad registration, exposure ranking and charging remain the same as for existing Powerlink operations. Clicks on the AI summary area are billed normally and included in the standard advertising click metrics.
Several regulated or review-sensitive categories are excluded from AI summary exposure, including hospitals and clinics, pharmaceuticals, medical devices, health functional foods, and finance and insurance. Naver attributes these exclusions to creative-review considerations. For eligible advertisers, the feature is provided automatically without a separate application unless they subsequently switch it off.
Landing-page optimization is becoming ad-copy optimization
The practical consequence for paid-search teams is that landing-page governance can no longer be isolated from creative governance. If the page contains stale pricing, vague product descriptions, outdated benefits or poorly qualified claims, those problems may affect more than conversion after the click. They can become source material for an AI system generating language before the click.
Conversely, a precise page gives the model cleaner material from which to summarize the offer. This does not mean advertisers should rewrite landing pages to manipulate AI-generated copy at the expense of users. It means that clarity, factual accuracy and semantic specificity now have another downstream consumer: the advertising model itself.
The trend is becoming difficult to dismiss as a platform-specific experiment. OpenAI can use landing pages to suggest ad creative; Google is connecting queries, generated assets and destinations more tightly inside AI Max; and Naver is now preparing to generate live Powerlink summary text from the advertiser’s own web content. In each case, the website is becoming part of the AI advertising control plane rather than merely the final URL.
The October 13 report will make the experiment measurable
Naver’s September 29 test will provide the first expanded desktop-and-mobile deployment before advertisers receive the opt-out control on October 7. The more revealing milestone may arrive on October 13, when formal deployment and seven-day generated-text reporting begin together. That is when advertisers can start comparing what they wrote on their pages with what Naver’s AI chose to tell searchers.
For performance marketers, this creates a new feedback loop: query intent influences generated copy, generated copy influences what the user sees, and the advertiser can later inspect those messages and improve the source information feeding the system. Paid search is becoming less about assembling a finite library of approved sentences and more about managing the information from which machines construct those sentences. Naver’s AI summary text makes that transition unusually explicit.