The AEO Conf 2026 decoded: what the data actually means for your brand, and what to do about it.
The shift is already happening — and it's bigger than you think
We've been tracking AI search behavior closely at Afluo while building Visibly AI, and the numbers from AEO Conf 2026 confirmed what we've been seeing in our own research: this isn't a slow-moving trend. It's a structural transformation happening in real time.
AI search is now 34% the size of traditional Google search by monthly session volume — and it's growing at +300% year over year. But the most striking stat isn't the size. It's where the usage is happening: 75% of all AI usage is on mobile apps. The explosion everyone was tracking on desktop was only a fraction of the real story.
Meanwhile, Google itself is changing faster than most marketers expected. AI Overviews are now shown in roughly 30% of all Google searches, surpassing featured snippets, which have been declining. Search is no longer a list of ten blue links. It's becoming an AI-first experience, and that shift is only accelerating.
ChatGPT dominates today — but the race is wide open
Right now, ChatGPT owns approximately 86% of the AI search market by monthly sessions in the US. That's a dominant position. But Gemini, Claude, Perplexity, and Grok are all growing fast, and the session charts from the conference show the gap is closing. Anyone who builds their strategy around a single AI platform is making a significant bet on a landscape that's still very much in motion.
The smarter play is to build visibility across all of them — which means understanding what makes AI models recommend you in the first place.
People aren't just searching with AI — they're doing and expressing
This is one of the most underappreciated insights from the conference. Based on OpenAI and Harvard research analyzing 1.1 million prompts, AI usage breaks down into three modes: 52% Asking, 35% Doing, and 14% Expressing.
The Asking category is what most people think about (traditional search-like queries). But "Doing" is where the most exciting growth is happening. As AI agents get better and integrations multiply, the portion of people using AI to actually complete tasks (booking, buying, comparing, executing) is going to grow dramatically. And "Expressing", creative, self-articulation use cases, is a reminder that AI has become a platform for communication, not just information retrieval.
For brands, this means your AI visibility strategy can't be limited to informational queries. You need to be present when people are in decision and action mode too.
The prompt is nothing like the search query
This is perhaps the most actionable insight from the entire conference. The average Google search is about 3.5 words. The average AI prompt is about 60+ words. And the specificity is completely different.
Someone Googling "running shoes" is browsing ten results before deciding. Someone prompting an AI with "best running shoes under $150 for a 45-year-old beginner runner training for a first half-marathon" is acting on one recommendation.
The data backs this up hard: 60% of ChatGPT prompts are 10 words or longer, with the largest bucket being 21+ words. Compare that to Google, where 2–3 word head terms drive the majority of searches. In AI, the long tail isn't a "nice to have", it's the main event.
This has direct implications for your content strategy. If your website only covers broad topics optimized for short head-term searches, you're essentially invisible to the most high-intent AI users. Your content needs to go deep — specific use cases, specific configurations, specific comparisons — because that's exactly what AI models are trained to surface when someone asks a highly specific question. This is something we see consistently in Visibly AI's prompt data: generic content gets ignored, and niche expertise gets cited.
Citations are fragmented — and Reddit alone won't save you
One of the most revealing slides from the conference showed the distribution of domains cited by AI models. The curve is extremely long-tailed. Even the most-cited domain (Reddit, at 2.36%) only accounts for a tiny fraction of total citations. YouTube, Wikipedia, LinkedIn, they all top the list but none of them are dominant.
What this means in practice: AI models are pulling from an enormous range of sources. Being mentioned on Reddit or Wikipedia helps, but it's not sufficient. The models are giving significant authority to hyper-specialized websites, niche publications, real customer reviews, and focused resources that don't show up in any traditional authority ranking.
We see this in our own research at Visibly AI too. A well-structured resource on a focused topic on a smaller domain can consistently outperform mentions from major platforms, because AI models are optimizing for relevance and specificity, not just domain authority.
What this means for your brand
The key insight from OpenAI's Chief Revenue Officer Ashley Kramer at the conference sums it up well: AI agents will increasingly be the ones choosing which companies, products, and services to recommend — and they'll do it based on highly targeted, quality content.
Here's what that translates to practically:
Being visible in AI search requires a fundamentally different approach than traditional SEO. You need to think about the quality and specificity of what AI models can learn about you, not just whether you rank for a few head terms. You need to be present across a diverse ecosystem of sources, not just the top-five platforms. And you need to understand which prompts you're winning and which ones you're losing — because unlike a Google ranking, you can't just look it up.
Measure it before you optimize it
This is exactly why we built Visibly AI. Most brands have no idea how they appear (or don't appear) when AI models are answering questions about their category. Are you the one being recommended when a potential customer asks ChatGPT for a solution or a product like yours? What about Perplexity, Gemini, or Google’s AI Overview?
Visibly AI gives you an AI Visibility Score based on thousands of real prompts run across the major LLMs. You get an accurate picture of when and where your brand and products get mentioned, which competitors are beating you on which queries, and — critically — what to do about it. We surface opportunities that no one else has spotted yet, because we're tracking the long tail of prompts where decisions are actually being made.
The shift toward AI-first discovery is real, it's fast, and the brands that build visibility now will have a significant advantage over those who wait until it's obvious.
Want early access? Head to getvisibly.ai to get on the list.
Data sourced from AEO Conf 2026, Similarweb, AirOps, and Ahrefs. Visibly AI research ongoing.
