Marketing Strategies in the AI ​​Search Era: 8 Questions for GEO, AEO, and SEO

In the AI ​​era, marketing becomes a matter of 'design' rather than 'execution'

In my last post, I shared my review of Zero Click SF , a meetup hosted by the U.S. AEO/GEO startup Profound . Today, I would like to share eight insights I gained from the meetup.

AI Search is not just changing search channels, but changing the very way marketing works.

While people remain important customers, AI agents have now become crucial as well. Therefore, marketing should no longer be limited to simply 'creating and distributing good content (web pages).' It must now be designed as a single system that encompasses brand information, context, trust signals, and conversion flows. 

In the changing era of AI search, what questions should marketers ask now? 

---

1. Who are the important customers in the AI ​​era? AI agents are also customers.

Before the AI ​​era, humans were the primary decision-makers.

• I entered a search term → viewed the search results → clicked the page (made a decision) → and decided whether to purchase on that page.

However, the flow has changed now.

• If you ask a specific question in natural language → the AI ​​reads, compares, and summarizes multiple sources → presents an answer first (judgment) → the user decides their next action based on that answer.

AI Agent: An entity that uses the web more actively than humans

In the future, AI will be the primary user of the web. This is because while humans have limitations in their search capabilities, agents can process hundreds or thousands of data points simultaneously without getting exhausted.

While the final decision-maker will still be human, AI will have authority over the entire journey leading up to it, from searching for information and discovering brands to assisting with purchasing decisions.

Therefore, AI should no longer function merely as a new marketing channel, but must be treated as an important customer and a quasi-decision maker.

2. In the AI ​​Era, Have Funnels Really Disappeared?

Some people say that “funnels have disappeared in the age of AI.” However, I believe that the interpretation that “the front end of the funnel is being processed like a pipeline within AI” is more accurate than the expression that “funnels have disappeared.”

This is because most of the 'Search – Awareness – Comparison – Consideration – Purchase' funnel occurs not outside the search results page, but within the process where the AI ​​generates the answer.

Therefore, to satisfy both types of customers (humans and agents) in the AI ​​era , you must target the AI's decision-making process and that funnel.

Differences between Traditional Funnels and AI Era Funnels

	quest	recognition	consideration	conversion	Purchasing/Onboarding
AS-IS traditional search environment	Enter search term	Check search results list	Compare after clicking on multiple pages	Persuading through website, sales, and product experience	Product experience starts after conversion
To-BE AI Search Environment	Input natural language questions	AI organizes brands/categories within the answers	AI compares, summarizes, and forms shortlists	Persuasion in landing and product experience following AI recommendation	Product wow is becoming more important, and AI partially absorbs pre-conversion information judgment

3. Is Product Led Growth (PLG) Still Valid in the AI ​​Era?

Traditional PLGs must now evolve as well. This was inspired by the insights of Reema Batta, VP of Growth Marketing at Figma. Citing the latest feature “Figma make,” Reema argues that there are clearly zones where it is difficult to grow using PLGs alone

Why is that? It is because products with different utility values ​​for each target persona are difficult to explain using a single funnel. Let's take Figma as an example. Founders may view it as a tool that facilitates rapid team collaboration and brand/product creation, PMs and POs evaluate it from the perspective of the planning-design-development handoff, and designers view it as a system and collaboration tool.

Examples of Persona-Specific Funnel Scenarios and Content Strategies – Figma Examples

Therefore, even for the same product, you must set different questions and comparison criteria for each persona and build different funnel scenarios for each persona.

Ultimately, the role of content is to provide context tailored to each persona's questions so that AI cites us as a recommendable service, while the role of the product is to design it so that users who went through the trouble of finding us can quickly experience "wow.".

In the AI ​​era, PLG alone is insufficient; you must design the product/onboarding segment where PLG operates strongly, along with the segment prior to that where AI compares and recommends.

4. How Should Persona-Specific Funnels Be Designed?

Even for the same product, the questions asked vary completely depending on the persona. AI platforms do not operate along a single path like traditional funnels; instead, they simultaneously reference communities, forums, documentation, and review sites. Every question can serve as the first step in the potential customer funnel.

⭐️ Funnel Building Strategy Template

Content strategy scenarios by funnel now need to be further refined so that they can be built for each persona.

	quest	recognition	Comparison/Consideration	verification	Conversion/Onboarding	Expansion/Utilization
Customer status	Recognizing the problem and seeking to solve it	Learning categories	Compare alternatives	Eliminate risk	Experience the value of the product	Expansion/repurchase at the team/organization level, etc
Nature of customer questions (prompts)	“How do I solve this?”	“What tools do you usually use?”	“What’s the difference between A, B, and C?”	"Can we actually do it, too?"	“What can I do right away after signing up?”	Can you continue to use it even if the team grows?
Keyword/Prompt Type	Problem type, situation type, JTBD type	Categorical, alternative search	Comparative type, substitute type, selection criterion type	Possibility verification type, counter-argument type, practical type	Starter type, Setup type, Fast results type	Operational, collaborative, scalable
Required content assets	Problem-solving articles, introductory guides, TL;DR	Category descriptions, use case posts, 'who uses it and why' content	Comparison table, comparison article, comparison page by persona	FAQ, Case Studies, Reviews, Implementation Checklist, Practical Q&A	Templates, demos, onboarding guides, first action calls	Collaboration Guide, Permissions/System Documentation, Advanced Usage Content
Brand Core Objectives	early discovery	Entering the category	Enter shortlist	Resolving doubts and building trust	activation	retention·expansion
Brand Representative KPI	AI citations, exposure, non-branded traffic	Category query exposure, brand mentions	Includes comparison keyword traffic and AI comparison answers	Time spent, FAQ traffic, conversion rate	Sign-up → First Value Reach Rate	Active users, extended usage

5. Which channels does AI cite more? : Media, Community (UGC), Review Platforms

Onsite alone is not enough; the importance of Offsite GEO.

AI also examines whether the description our brand defines itself is repeated in third-party channels using similar language and context. Therefore, in the era of AI Search, brand trust is not built solely within a brand website, but signals accumulated across the entire web.

If Onsite GEO is about organizing a brand's official description, Offsite GEO is closer to designing "how others describe us.".

Onsite and Offsite GEO

• Onsite GEO: This refers to GEO activities performed on websites, blogs, product pages, etc., owned by the brand. It is closer to the realm of clearly defining the brand entity and explaining the brand in a way that is easy for AI to understand.
• Offsite GEO: This refers to GEO formed on third-party channels not owned by the brand. The key lies in "what language others repeat about us" rather than "how we talk about ourselves." Representative offsite channels media, review platforms, communities (UGC), social media, and other web documents.

North America Offsite Channel Strategy – Reddit, G2, LinkedIn, and Media Outlets

The importance of the Offsite role is already a very significant fact in the GEO industry.

According to data released by Similarweb, the most cited web pages in AI responses were 1) News & Publishers (Media) and 2) Review & UGC. While citation rates vary by AI model, these two channels can be considered the most trusted across all web pages.

In particular, in the North American market, channels such as Reddit as a community (UGC platform), G2 as a review channel, and LinkedIn as a B2B social platform are fulfilling these roles.

In particular, in the North American market, channels such as Reddit as a community (UGC platform), G2 as a review channel, and LinkedIn as a B2B social platform are fulfilling these roles.

---

What is Korea's Offsite strategy?

Korea is not a market like the U.S. where a single platform like Reddit or G2 operates strongly. Instead, it is closer to a market where various channels such as Instagram, Threads, LinkedIn, industry communities, media, and personal blogs share that role.

In other words, rather than asking “which is the best channel,” the Offsite strategy should first examine which external signals work most persuasively when our target customers ask a certain question

Review Sites vs. UGC vs. Media: Which Is More Important?

Then, which is more important among review sites, UGC, and media? This depends on the nature of the question, rather than the superiority of the channels themselves.

• Reviews and comparative content are strong for comparison and multiple-choice questions. Example: “Which tool is better, A or B?”
• Communities and UGC are strong for questions regarding real-world usage and practical application. Ex: “How is it actually used in the field?”
• Media and expert contributions are strong in contextualized questions regarding definition, authority, and the market. Example: “Why has this category become important?”
• Business social networks like LinkedIn are advantageous for expertise- and people-centered, trust-based questions. Example: “Who actually holds what perspective in this field?”

Offsite AI Citation Roles by Channel

Offsite types	Role (Layer)	Referenced prompt type
Media (News, Research, Industry Media, Conference Content)	Authority layer (definition of reliability and industry standards)	Definitional question (“What is AEO?”), Industry description (“AI search market structure”), Trend question (“Recent GEO strategies”), Thought leadership question
Review (Review platform, comparison article)	Experience layer (provides grounds for selection judgment)	Recommendation questions (“best CRM tools”), purchase review questions (“Is HubSpot worth trying?”), Shortlist creation questions
Community (Reddit, Quora, online community)	Experience + relevance layer (real-world usage context/field signals)	Comparative questions (“Webflow vs Framer”), experience questions (“What is it like in practice?”), problem-solving questions (“Why is it slow?”)
Social (LinkedIn, X, YouTube, Medium, etc.)	Freshness + relevance layer (latestness · expert context)	Questions on the latest trends (“How has SEO changed these days?”), questions on expert opinions (“How are marketers responding?”), and questions on conferences and issues
Web Directory	Decision layer (provides criteria for compressing and comparing purchase candidates)	Recommendation-type questions (“B2B SaaS recommendations”), category exploration questions (“What project management tools do you have?”), vendor shortlist questions

① Definitional questions are media-centered

For a brand to appear as a “source for concept explanation,” mentions in industry media are key. Large media outlets often block AI crawlers, so their overall citation rate tends to be very low. In such cases, media specializing in the relevant category are practically more effective.

② Recommendation-type questions are Review + Directory-centric

This layer has the greatest influence on actually making it onto the AI ​​recommendation list. In the case of the software review platform G2, there is an instance where citations rose by up to 40% for content written in first-person language.

③ Comparison and experience questions are overwhelmingly influenced by the community

Reddit-style UGC is currently a citation source repeatedly referenced by almost all AI models. Brands should not try to “own” the community, but rather approach it by being “invited” to it.

6. AEO/GEO Content Strategy: 'Write to be cited better,' not 'Write more.'

In the age of AI, it might seem advantageous to create a lot of content, but in reality, that is not the case. Google also clearly explains that there is no separate strategy for being cited in AI Overviews and AI Mode; instead, basic SEO principles, text accessibility, structuring, and trustworthy content remain important.

AEO/GEO Content Strategy How to

We need to go a step further than simply structuring and writing well. Instead of writing long texts, we must now write in a format that AI can immediately retrieve as an answer to a question

In practice, these things become important.

• Rewrite existing content and update with the latest information
• Place a TL;DR (conclusion summary) or context summary at the top of a product page or blog
• Insert FAQ, Comparison Table, and Q&A Structure
• Naturally incorporating reviews or customers' actual language into the text
• Creating a Q&A page structure

Content Strategy to Be Cited in the Top 3 AI Recommended Products

One of the most significant changes in the AI ​​search environment is that AI generates a shortlist of candidates before users can directly 'explore' brands . To be included in this shortlist, evidence is required to answer the following questions, going beyond mere "high exposure."

• In what category/alternative group does AI classify us ?
• Are the comparison criteria (selection criteria) clear? (Price/Difficulty of implementation/Suitable situations/Unsuitable situations)
• basis for trust? (Case studies, reviews, third-party mentions, data)
• Can it be explained with a consistent one or two-sentence summary even if the question changes ?

Conversely, it is also clear what should not be done.

AEO/GEO Content Strategy Don't

• blindly churning out mass-generated content
• It has many emotional sentences but a weak information structure
• Focusing only on a visually appealing UI with meager text information
• There are only brand claims and no signs of third-party verification

Ultimately, citations stem not from quantity, but from recency, structure, reliability, and expertise

7. Agentic Commerce, What Is the Future? From Recommendations to Purchases.

Now, AI is not just going as far as saying “I recommend this product,” but is preparing to compare products and even assist with the purchase process.

OpenAI launched an Instant Checkout feature in 2025 that allowed agents to directly analyze prices, inventory, and reviews on the web and complete the final payment, but it was discontinued in March 2026. However, due to issues such as the accuracy of price and inventory data, the user experience did not meet expectations. Although OpenAI's first attempt failed, important standard infrastructure is being established in the process.

Agentic Commerce Protocol (ACP): An open protocol jointly developed by Open AI and Stripe. It is an interface standard that enables AI agents to directly interact with external commerce systems (shopping malls, payment, inventory APIs, etc.) and autonomously handle product search, comparison, purchase, and tracking.

Universal Commerce Protocol (UCP): A universal transaction protocol jointly developed by Google and Shopify. Its key feature is that it is not tied to a specific platform, enabling any AI agent to communicate with any commerce platform in a standardized manner. as the foundation for agents to compare multiple shopping malls simultaneously and execute optimal transactions.

What Should Brands Prepare for in the Era of Agencyic Commerce?

Brands must become interface data that assists AI in purchasing decisions.

Moving forward, brands must not simply create "human-friendly pages." AI must be able to analyze and evaluate price, stock, options, reviews, refunds, shipping, and seller credibility. In other words, future detail pages and product data must go beyond SEO and become interface data designed to assist AI in making purchasing decisions

And this applies directly to B2B as well. Even if you don't purchase enterprise software directly within ChatGPT today, AI is already deeply involved in pre-purchase stages such as shortlisting, supplier comparison, requirements organization, and internal summaries. Therefore, what brands need to prepare for going forward is not a competition for exposure, but a structure where AI can compare, explain, and purchase on their behalf.

8. So, what should marketers do going forward?

The Era of Marketing Engineering

Thanks to AI, marketing is entering a stage where it can be designed into a system that monitors, diagnoses, executes, and iteratively improves. This is sometimes referred to as the era of "marketing engineering.".

In the era of marketing engineering, the capabilities of marketers must also evolve.

What is Marketing Engineering?

• The work of designing a system so that data automatically finds the answer and drives performance growth
• Detect and determine via system/agent -> Execute -> Measure -> Reflect the results back into system improvement</aside>

The Role of the Marketer in the Era of Marketing Engineering

While the focus previously was on channel-specific execution, copywriting sense, and campaign management, the ability to orchestrate various tools, agents, data, content, products, and sales as a unified wholebecoming more important.

In fact, even in the SEO era, writing well alone did not guarantee results. Content, information structure, technical SEO, internal links, conversion flow, and web experience had to work together. In the era of AI Search, this aspect is becoming much more pronounced.

Now, marketers go beyond merely planning and producing content

• Organize the brand context and
• Connecting signals that AI can refer to and
• Predicting what answer will be generated from a given question and
• You must design a structure that enables verification and conversion after the recommendation.

From this perspective, I believe the capabilities of marketers with an engineering mindset will grow within the organization

---

Conclusion of the 8 questions.

• Define which customers and in what context our brand should appear, and
• Directing product experience, brand context, and storytelling across all digital channels.
• Ultimately, building trust capital that both humans and AI can agree with.

Even as channels change, technologies evolve, and protocols emerge, these three things remain constant. In the age of AI, the brands ultimately chosen are those that are accurately understood and trusted by both people and AI.

Elifunt wants to be the marketing team that does this job best. Rather than feeling like we are working with an agency, we want to be a team that deeply understands our brand's context and designs together.