In 2024, generative AI as a whole ranked just 11th among information-search platforms. By the first half of 2026, ChatGPT alone had climbed to 4th. That's the figure at the center of the "First Half 2026 Marketing Trend Report" released this year by Incross, an integrated marketing firm under the SK Networks group. What stands out even more than the number is the speed. The platforms currently holding the 1st through 3rd spots took anywhere from several years to more than a decade to get there. ChatGPT broke into that tier in under two years.

What this shift means for the people who create and distribute content doesn't seem to have been fully discussed yet.

73.8% of Users Open AI to Search First

According to the Incross report, 73.8% of generative AI users named information search as their top use case for AI. Translation and summarization came next, with text generation further down the list. Most people are reaching for AI as a way to find information before they think of it as a writing assistant.

That number matters for content creators because traditional search and AI search deliver information through fundamentally different structures. Type a keyword into Google or Naver (Korea's dominant search portal) and you get a list of results; you click the link you want and land on the original page. Ask an AI a question, and it synthesizes the relevant information into a single answer. Content reaches the user without a click ever happening.

That structural difference matters enormously from a creator's standpoint. When your article ranks at the top of Google's results, a reader clicks through, reads the original piece, and remembers your name and your channel. When an AI summarizes your content to answer someone else's question, the substance gets through, but the source often doesn't travel with it. No traffic reaches your site, and readers have fewer chances to ever register you as the author.

To be fair, AI systems are moving toward including source links in their answers, and some already display citations. But at this point, it's hard to say that exposure through AI search functions the same way exposure through conventional search does.

Two Different Optimizations, Two Different Rulebooks

For the past decade, the baseline for online content strategy has been search engine optimization, or SEO. Creators learned how Google's and Naver's algorithms evaluate pages, wove in keywords naturally, refined heading structures, and built up backlinks to boost visibility. Whether that same formula holds up in an AI-search environment is far from settled.

There's real overlap between the two forms of optimization. Concrete, fact-based content tends to do well both in search rankings and in AI citations. But there are also real divergences. Search engines weight technical signals heavily — metadata, inbound and outbound link counts, page load speed. AI systems tend to care less about those technical signals and more about the density of the content itself: does it directly answer a specific question, does it cite numbers and sources, and does it offer a perspective distinct from the dozen other pieces written on the same topic.

A sentence like "this trend deserves attention" is far less likely to surface in an AI answer than one like "according to the Incross report, 73.8% of generative AI users use AI primarily to search for information." Numbers and sources give an AI something concrete to cite; generalities and abstractions don't.

This is exactly where a "well-optimized SEO piece" can backfire. An article that hits keyword density by repeating similar phrases without adding concrete information, or one with a clickbait headline sitting atop shallow content, can still score well on SEO — but it gives an AI no reason to cite it. As AI search grows as a share of how people find information, that kind of content stands to lose ground relatively — which is the paradox the headline points to.

But There's a Hole in This Analysis

Still, it's worth slowing down here. The entire conversation around "content strategy optimized for AI search" remains largely unproven.

Companies don't clearly disclose the criteria ChatGPT, Gemini, or Claude use to select and cite web content. Citation behavior shifts with every model update, and results differ depending on whether a user has real-time search enabled. Much of the "AI SEO" advice circulating right now rests more on logical inference than on hard empirical data.

Zooming out further: even with ChatGPT at 4th place among search platforms, the top three spots still belong to established players. AI search is growing fast, but at this point it may be more accurate to describe it as one more channel added to the mix rather than a replacement for existing ones. Declaring that "SEO is dead" or that creators need to tear up their entire strategy for AI search is premature right now.

I'd rather treat this shift not as "a new optimization technique to master" but as an opportunity to take a fresh look at where and how readers actually find information.

What You Can Adjust Right Now

Responding to this shift doesn't have to mean abandoning what you're already doing. There's room to adjust within your existing approach.

Start by asking, every time you write, exactly which specific question this piece answers for someone. A broad subject like "the future of AI marketing" is far less likely to show up in an AI answer than something like "what to check first when your newsletter open rate drops below 20%" — a piece that directly answers a specific question in a specific situation. This doesn't conflict with SEO principles at all. Content that answers the questions readers are actually asking is more likely to get picked up by both search engines and AI, which means the two directions point the same way here.

It's also worth reviewing whether your writing cites numbers and sources with enough specificity. "Experts are paying attention" builds far less trust — and gives AI far less to cite — than "according to Organization A's 2026 report, the figure is B%." If you're already writing this way, you're already ahead. If not, this is a change you can make right now at no added cost.

There's another angle worth considering. As AI increasingly summarizes and delivers your content on your behalf, maintaining a relationship with readers through web traffic alone gets harder. Readers connected to you directly — through an email newsletter, a Telegram channel, a YouTube channel — without going through a platform's algorithm are insulated from however AI citation practices evolve. The people who keep performing well amid fast-changing conditions tend to share one thing: an asset that keeps working no matter how circumstances shift. For a content creator, a directly connected reader base is about as close to that asset as it gets.

It took two years to go from 11th place in 2024 to 4th in 2026. Predicting the ranking two years from now is hard. But figuring out where the starting point of readers' information search is heading right now isn't.