Hello, I am Ji-eun Park, Content Growth Marketer at Elifunt.
It has already been three weeks since I joined Elifunt. I thought it was about time to introduce myself through the blog, but I am back a little sooner than expected with a review of the GEO Roadmap 2026
Although I joined the event preparations midway and spent less time with the team than others, that actually allowed me to experience the event from a similar perspective to the attendees. I would like to share with you what was discussed, what particularly left a lasting impression on me at the event, and what I personally saw, heard, and felt.
Designing a GEO Strategy for Businesses in the Era of AI Search
How often and in what context is our brand being mentioned in AI search results? Nowadays, instead of comparing search results one by one, customers first ask AIs like ChatGPT, Gemini, and Perplexity. Then, within the answers organized by the AI, they discover brands, compare them, and decide their next course of action.
We are now in the full-fledged era of AI search. Companies are facing a new question: “What should we do to ensure our brand is cited in the answer when a customer asks an AI?” To collectively brainstorm the answer to this question, Elifunt held the GEO Roadmap 2026 seminar on July 2nd
Seeing all tickets sell out in just two weeks after announcing the event , we were able to confirm that many companies share the same concerns as we enter the era of AI search. In particular, thanks to the diverse range of companies applying to participate regardless of industry or size, it was an occasion where we realized once again that everyone is starting from the same starting line in the new environment of GEO.
Prior to the event, we collected questions in advance to help resolve as many concerns as possible from the applicants. Nearly half of the applicants submitted questions. We selected 12 frequently mentioned topics from among them, obtained answers directly from the speakers, and created question cards.
Top 3 Question Cards Most Picked by GEO Roadmap 2026 Attendees
TOP 1. I have been running content on a Naver blog; what should I improve from a GEO perspective?
Global AIs like ChatGPT and Gemini find it difficult to reliably read and cite posts on Naver blogs. This is because Naver restricts access for the purpose of training or search enhancement by global AIs.
Therefore, if you have operated many Naver blogs, you do not need to discard those assets, but you should divide your roles. It is recommended to maintain Naver blogs as a point of contact with Naver Search and domestic users, while restructuring core content on an official website or resource hub. You must organize
everything from FAQs for specific customer questions, service definitions, comparison criteria, case studies, technical justifications, and author and company information into web pages accessible to AI. From a GEO perspective, supplementing your content is not about writing more blog posts, but about building a database of reliable answers in places where AI can read them.
TOP 2. To what extent do external mentions, such as product reviews, communities, and YouTube, influence the AI's response?
External channels that have built trust through search have a significant impact. Among these, media relations are particularly effective in boosting exposure in the short term, as AI actively utilizes them when searching for the latest information via web browsing.
Similar to traditional SEO, AI actively cites channels with content rich in the context of actual experiences and interactions. For instance, regarding hospitals, while university hospitals overwhelmingly dominate in terms of official information and health knowledge, direct questions such as "recommend a hospital for a specific disease" tend to prioritize recommendations for hospitals that have been well-marketed on blogs or review service channels. When regional context is added, review content that incorporates geographical information receives even better exposure. Ultimately, the key variable for AI response exposure is "on which channels and within what experiential context we are being mentioned" outside of the website.
TOP 3. Since each AI provides a different answer, how should I monitor and report it?
There is no single solution that satisfies all AI searches. Each AI, such as ChatGPT, Perplexity, and Gemini, has different preferred sources and response styles, and these differences are becoming increasingly distinct. Therefore, while monitoring the current status, it is advisable to prioritize the core AI touchpoints most frequently used by your customers.
If resources are limited, it is more realistic to focus on improving performance in one key AI search and sharing the results with the team, rather than attempting to optimize all AI searches simultaneously from the start. You can then expand the scope of optimization to weaker AI search channels. Regarding monitoring, it is also efficient to first define the goals the company aims to achieve through GEO and then consistently observe a single key metric connected to those goals.
We plan to cover all the questions and answers in detail in future content. Please look forward to it!
GEO Roadmap 2026 Session Highlights
The GEO Roadmap 2026 seminar did not stop at introducing the concept of GEO, but covered brand entity design, off-site trust signals, on-site website structuring, and performance measurement as a single flow. At the event, rather than asking "How will it appear in AI responses?", "What needs to be prepared to become a knowledge source that AI can trust and refer to?"was repeatedly raised.
We have summarized the key contents of each session from Elifunt's perspective, featuring a total of six experts ranging from CEO Yeji Kim of Elifunt, Division Head Cho-rong Han of Ascent AI (Listening Mind), Division Head Gong In-hee of Ham Shout Global, Team Leader Chan-ki Hong of Daehong Communications, CEO Seong-gyu Lee of Blue Dot AI, and CEO Jae-cheol Lee of HaloX.
Session 1. GEO Overview: GEO is an extension, not a replacement, of SEO.
CEO Kim Ye-ji began the first session with 'Why GEO Is Needed Now.' While SEO is still necessary, GEO is not complete with SEO alone. If SEO is about making search engines discover and understand our website, GEO is closer to managing how AI interprets our brand, including the categories, evidence, and contexts used.
AI responses begin with the customer's question. Depending on the question the customer asks, the information the AI collects, the brands it compares, and the final answer all vary. Therefore, GEO's starting point is not "I want to appear in the AI response," but rather verifying "what questions customers ask and whether we have reliable sources to answer them.".
The presentation also shared the current status of GEO responses by 548 domestic companies. The overall average score was below 50 points, with on-site performance at 54 points and off-site performance at 37 points. Only 6% of companies could be considered to be in the excellent range. This implies that most companies are still on a similar starting line.
Check the current status of our companies based on industry and growth scale through the GEO reports of 548 domestic companies presented in CEO Yeji Kim's session
Session 2. Brand Entity: Brands chosen by AI are not created by chance.
at Ascent AI (Listening Mind), introduced a method for identifying the context of consumer questions from search intent data and connecting it to GEO strategies. At the heart of the presentation the Category Entry Point (CEP). CEP refers to the situation and context in which consumers recall a specific category or brand.
The form of search is shifting from short keywords to long prompts. Within these longer questions, you must identify the Customer Experience Points (CEP) to preempt based on the customer's needs, selection criteria, grounds for trust, and emotional state. It requires a process of reading market changes through search data, identifying which products consumers seek in what situations, and then converting this into a set of prompts that need to be managed. The important thing is not to look only at brand-name questions. You must examine non-brand questions, comparison questions, and review questions as well to understand the context in which AI invokes our brand.
Brands chosen by AI are not created by chance. You must identify customer questions, structure the context within those questions, and consistently manage the evidence that AI can understand. GEO is not a one-time campaign, but rather closer to a brand operating system.
Session 3. Offsite GEO: Trust built outside of search leads to AI answers.
Ms. Gong In-hee, Head of Division at Ham Shout Global , explained Off-Site GEO, focusing on trust signals generated outside the website. While many companies immediately think of website redesigns when discussing GEO, AI does not rely solely on its own website. It also references information accumulated externally, such as media, social media, communities, reviews, and influencer content.
In particular, many people showed interest in the examples of Naver AI Briefing and Meta's AI Mode. Naver announced that it would actively support creators cited in AI Briefing, while Meta is moving toward utilizing content from Facebook, Instagram, and Threads in its AI responses. Going forward, the voiceovers, subtitles, and captions on Instagram Reels can also serve as brand information assets. Emotional content alone is no longer sufficient. Include information that AI can read and quote!
Offsite GEO's conclusion is that we should first look at " which channels AI actually references in our industry" rather than "." Trusted channels vary by industry. Magazine websites may have a greater impact on lifestyle brands, company websites and Naver blogs on health food products, and professional influencer content on specific consumer goods.
Session 4. Onsite GEO: We need pages that AI can understand, rather than pretty pages.
For CEO Yeji Kim's fourth session, we returned to on-site.
The era of pretty detail pages is over. While making things look good is still important, structuring them so that AI and search engines can understand them has now become more critical.
On-site is a channel that a brand can directly control. At the same time, the point where the final conversion occurs. Customers can enter our website only if the brand is mentioned in AI responses or cited via links, and once they arrive, service pages, case study pages, and content pages must persuade them.
CEO Kim Ye-ji pointed out the limitations of image-only detail pages, one-page websites, and supplier-centric language. Pages composed solely of images may look aesthetically pleasing to humans, but they lack sufficient text for AI to read. Conversely, pages equipped with titles, descriptions, H tags, body text, and schema markup allow AI to more clearly understand what kind of company it is, what services it provides, and what customer problems it solves.
Schema markup was also an important topic. A schema is like a business card you hand to AI and search engines. You can specify the meaning of a page through structured data such as Organization, Service, Product, FAQPage, Article, and Person. However, it is dangerous to include content in the schema that is not actually on the page or to repeat it excessively. An incorrect schema can pose a greater risk than having nothing at all.
Content strategy was also covered in the fourth session . The necessary content varies depending on whether the customer is unaware of the problem, knows the problem but does not know the solution, or is already comparing vendors. Elifunt categorizes this into inflow content, conversion content, and core content. In the era of AI, the value of core content containing a company's unique experience and expertise is also growing.
Session 5. GEO Performance Measurement: KPIs are Visible Only When the Purpose Is Clear.
from Daehong Communications discussed the differences in the purpose of viewing GEO. He pointed out that the most frequently encountered problem in the field is that the imperative—"Shouldn't we be doing GEO too?"—comes first, while "why we should do it" and "what we will prove"are pushed to the back burner.
GEO is not an activity where immediate results are guaranteed in the short term. However, if you start without a purpose, you will be asked, "What are the results?" after about three months. If you start looking for KPIs at that point, the project may stall, and the budget and momentum for execution could disappear.
Team Leader Hong Chan-ki explained that to perform measurement, you must first define the questions. Just as you wouldn't formulate a strategy in SEO without keywords, you shouldn't start in GEO without a prompt. At this point, the prompts and metrics to track will vary depending on whether you aim to increase brand awareness and positioning, connect AI traffic with sales, or manage reputational risk.
Ultimately, the first step in GEO performance measurement is not selecting metrics, but agreeing on business objectives. Only when the objectives are clear can KPIs be seen, and only when KPIs are seen can execution follow.
Session 6. GEO Performance: It must lead to sales.
BlueDotAI CEO Lee Seong-gyu posed the most realistic question regarding whether GEO performance can lead to revenue. If GEO metrics have improved, can we say that this has led to actual sales or conversions?
Two contrasting cases were presented during the presentation. For the green tea brand, domain citation rates increased following the implementation of GEO, and growth in purchase conversions and sales was also confirmed in GA4. Conversely, while the insurance brand saw improvements in GEO metrics and AI-related performance, the actual increase in consultation-based sign-ups did not meet expectations.
CEO Lee Seong-gyu proposed product type as the criterion for explaining this difference . He noted that GEO performance varies depending on whether the product is search good, where characteristics and quality can be compared before purchase; experience good, where quality is assessed through post-purchase experience; or trust good, where quality is difficult to judge even after purchase. For products closer to experience goods, such as green tea, AI search traffic is relatively more likely to lead to a purchase. On the other hand, for products with strong trust characteristics, such as insurance, the effect of being added to the consideration set may occur before leading to an immediate subscription.
Therefore, to explain GEO performance in terms of revenue, the correlations between GA4's AI Assistant channel, AI acquisition, purchase-consultation-lead data, and existing internal metrics. To convincingly demonstrate revenue contribution, GEO metrics must be linked to indicators that the company already trusts.
Session 7. From SEO to GEO: Measurement Workflows Must Be More Three-Dimensional.
HaloX CEO Lee Jae-cheol introduced a workflow for measuring performance by integrating SEO and GEO. While impressions, clicks, rankings, and CTR were key metrics in the SEO era, the GEO era requires considering citations, mentions, and share within AI responses as well
CEO Lee Jae-cheol explained the measurement framework “Read, Fetch, and Visit.” First, AI bots must be able to access our site. Checking only robots.txt is not enough. We need to verify whether GPT bots or Claude bots are blocked at the network level, and whether AI bots are unable to see rendered information due to the CSR structure.
And you need to look at the content that AI retrieves. For specific keywords, 50–70% of the top 10 SERP results tend to appear in the AI Overview. Ultimately, to dominate the AI Overview, you must analyze the SERP structure, domain characteristics, titles and descriptions, and topic clusters together.
Finally, you need to find the "People Also Ask" question in the keywords, expand it into a prompt, and then check which subquery GPT uses to navigate the answer. By connecting GA4's AI referrer, GSC's keywords and page data, Google Search's question data, and GPT's fan-out query, you can view SEO and GEO as a single measurement stream
Are you ready for our organization's GEO Roadmap?
Following a sell-out of tickets, the GEO Roadmap 2026 seminar concluded with great success as all applicants attended. It was a meaningful time where 100 industry experts and marketers with diverse concerns gathered to discuss the direction of the AI search era together.
In particular, many attendees shared feedback stating, “I was able to formulate a real action plan on what to check and improve in our service right now,” and “It was beneficial to learn specific practical checkpoints.” It was an even more meaningful session in that it went beyond simply introducing GEO and delivered insights that can be immediately applied to actual business operations.
There was one message that ran through this seminar. It is that GEO is not a technology that forces brand exposure into AI responses, but rather a process of understanding what questions customers actually ask, creating content and websites that best answer those questions, and building consistent signals of trust across various external channels.
While AI search is evolving rapidly, the criteria for brands ultimately remain the same. Brands that provide the most reliable answers to customers' questions first will be chosen by AI as well.
Are you wondering where to start with GEO? Elifunt, your organic marketing partner, will diagnose your company's current situation and work with you to develop GEO strategies and implementation plans tailored to the era of AI search.
