← Back to blog

AI search: Getting ready to be found on ChatGPT and Perplexity

800 million people now use ChatGPT every week. Three out of four Czechs aged 15–64 reach for AI on a regular basis. And only 1% of users click a source inside an AI answer. If the AI doesn't recommend you directly, your business simply doesn't exist in the eyes of a new generation of customers — and trust in that kind of recommendation runs deeper than trust in any ad.

Key takeaways

  • Over 800 million people use ChatGPT weekly, and three out of four Czechs aged 15–64 regularly use AI.
  • Only about 1% of users click a source inside an AI answer — what matters is being recommended directly in the text.
  • ASEO essentials: a website readable without JavaScript, structured data (schema.org), question-and-answer content, and consistent company data across the web.
  • An AI recommendation feels more trustworthy than advertising — whoever optimises first gains a disproportionate advantage.

Sam Altman put it bluntly at the Sequoia AI Ascent conference in May 2025: "Older people use ChatGPT as a replacement for Google. People in their twenties and thirties use it as a life advisor. And college students use it as an operating system." If you own a business and your customers fall into any of those buckets — and they do — then the question of whether AI will recommend you for the next five years is being decided right now. The rules look very different from classic SEO.


What's actually happening to search right now

In October 2025, ChatGPT crossed 800 million weekly active users (Sam Altman, OpenAI DevDay, October 6, 2025). For context: Facebook hit the same milestone in 4 years, Instagram in 2 months, ChatGPT in 5 days. OpenAI's ARR jumped from $2 billion in 2023 to over $20 billion in 2025 — a tenfold increase in 24 months (Sarah Friar, CFO of OpenAI, January 2026).

The competition isn't standing still. Perplexity AI, which founder Aravind Srinivas positions as "discovery of knowledge," grew from a $121 million valuation (April 2023) to $20 billion (September 2025). It now handles 780 million queries a month. Anthropic's Claude sits at roughly 30 million active users. And Google rolled out AI Overviews — generative summaries sitting directly above the organic results — which Sundar Pichai, on stage at Google I/O 2025, called "one of the most successful launches in Google's history in the last ten years."

In the Czech market the picture is just as dynamic. According to Marketing Miner data (SEO Restart, September 2025), the share of Google queries triggering AI Overviews in the Czech Republic climbed from 10.5% in May 2025 to 19.6% in September 2025. A STEM/MARK survey (January 2026) found that 75% of Czechs aged 15–64 use AI tools on a regular basis; 89% of people under 29, 78% in the 30–44 bracket, and roughly half of those over 44. Deloitte TMT Highlights for the Czech Republic (2026) reports that 45% of Czechs regularly use ChatGPT — three times more than in 2023.

At the same time, Gartner predicts that by the end of 2026, traditional search volume will drop by 25%. As Alan Antin, VP Analyst at Gartner, put it (February 19, 2024): "Generative AI is becoming a substitutive 'answer engine' that replaces queries that previously went to traditional search engines. This will force companies to rethink their marketing channel strategy."

The psychology of trust — why an AI recommendation hits differently than an ad

This is the heart of the shift. And for a small business, the key opportunity. Until recently, a customer searching for "best plumber in Prague" got a list of links. They knew the rankings were largely gameable — ads, SEO, sponsored content. They learned to tell ads from organic results, to weigh reviews, to compare options.

Today, that same customer types the same query into ChatGPT and gets back a single synthesized answer: "For plumbing services in Prague I recommend X — excellent reviews, fast response time, and experience with…" And they click X. Or they pick up the phone and call X directly. No comparison shopping. No double-checking. Without ever perceiving that answer as advertising.

Algorithmic authority: AI as a neutral expert

A 2026 paper in the academic journal Lex Wire Journal coined the term "authority hallucination" — users perceive an AI recommendation as a neutral synthesis of facts rather than paid placement. The principle of trust transfer says that trust in a platform carries over to whatever that platform recommends. If I trust ChatGPT, I automatically trust whatever ChatGPT recommends.

A key 2026 finding from AirOps: 85% of brand mentions in AI search results come from third-party pages, not the brand's own domain. That flips twenty years of SEO strategy on its head. Being strong on your own website is no longer enough — you need to be cited on Heuréka (the leading Czech price-comparison and review portal), on Wikipedia, on industry portals, in discussion forums.

Pew Research: only 1% click a source inside an AI answer

In July 2025, the Pew Research Center published a study based on a panel of 900 American adults and 68,879 real-world search queries. The findings are pivotal:

  • 72% of users find AI summaries useful (1 in 5 finds them extremely useful).
  • 54% trust the information in an AI summary at least somewhat.
  • When an AI summary is present, click-through on traditional links drops from 15% to 8% — nearly cut in half.
  • Clicks on a source inside an AI summary: 1%.

That last number is the key to understanding the new dynamic. The AI answer is the destination. It's not a gateway to a source — it's the final answer. As Pew summarized the underlying psychology: "Users don't decide whether to trust AI based on technical understanding. They build confidence through repeated positive experiences with AI-generated information, often without even realizing they're interacting with artificial intelligence."

Trust in AI doesn't come from rational deliberation — it's built by habit. And once it's built, it's hard to shake. Your customers are forming that habit right now.

Conversion rates 4.4× to 23× higher than traditional search

If the psychology isn't enough, the economics are. In June 2025, Semrush analyzed more than 500 topics across industries and found that visitors arriving from AI recommendations convert 4.4× better than visitors from organic Google. Ahrefs' internal data showed an even sharper gap: 0.5% of traffic from AI generated 12.1% of sign-ups — a 23× higher conversion rate.

Specific case studies:

  • Seer Interactive (multi-vertical): ChatGPT converts at 15.9%, Google organic at 1.76%. ChatGPT visitors read 2.3 pages per session vs. 1.2 for organic.
  • Visibility Labs (94 e-commerce brands, 12 months of GA4 data): ChatGPT traffic converted 31% better than non-branded organic. ChatGPT sessions grew 1,079% year over year.
  • Microsoft Clarity (study of 1,277 publishers, Q4 2025): referrals from Copilot convert 17× better than direct traffic.

The mechanism behind this asymmetry is what the analytics firm Lantern calls "pre-qualification": the AI has already done the comparing, the vetting, and the recommending before the visitor clicks. By the time they arrive at your site, they're not a shopper sifting through 10 links — they're a buyer who was told "go this way" by a source they trust as neutral.

The generational shift: ChatGPT as the new family advisor

If you sell to customers aged 40–50, you might be tempted to say: "This doesn't apply to me, my customers are older." That would be a mistake for two reasons.

First — the 40+ generation is catching up to younger users fast. Per the Czech STEM/MARK survey (January 2026), roughly half of Czechs over 44 now use AI regularly. Sam Altman confirms it: "Older people use ChatGPT as a replacement for Google." For your target audience, ChatGPT is functionally a "smarter Google" — and the trust logic that gets you recommended works exactly the same way it does for younger users.

Second — even if your core audience isn't yet abandoning traditional search, their children, grandchildren, and younger employees already are. And in a B2B context, that research is what decides who makes the shortlist. According to the 6sense Buyer Experience Report 2024, 80% of B2B buyers have a preferred vendor in mind before they ever reach out to you. 95% of B2B purchases start with online research. If that research is now happening in ChatGPT or Perplexity and your company isn't there, you don't make the shortlist.

A look ahead — where this is heading

This isn't a trend that will burn itself out. It's a structural shift. And the industry's leaders are saying so out loud:

Whoever wins the personal agent — that's the big thing, because you will never go to a search site, you will never go to a productivity site, you'll never go to Amazon again.

— Bill Gates, Goldman Sachs / SV Angel event, May 2023

In an April 2026 conversation with Stripe CEO Patrick Collison, Google CEO Sundar Pichai described the future of search as an "agent manager" — users will run "many threads at once" and complete tasks rather than scrolling through links. He called 2027 an inflection point for fully agentic enterprise workflows. Eric Schmidt, the former Google CEO, put it even more starkly at TED 2024: "The arrival of this intelligence is the most important thing that will happen in human society in the last 500 to 1,000 years. And it's happening in our lifetime. If you don't use this technology, you will not be relevant compared to your competitors."

What does that mean in practice? Your customer may never visit your website at all before choosing you. All it takes is for the AI agent to mention you in its answer. And if you aren't in that answer — because the AI simply doesn't see you — you don't exist. Aravind Srinivas, CEO of Perplexity, says it without sugar-coating: "We've been very transparent that this new interface is not going to send you as much traffic as before." Perplexity is deliberately built to break the ad-supported business model. And it's not alone.

Data from Cloudflare shows just how quickly the economics of content are changing. In 2015, Google scraped web pages 2 times for every link it sent back. By 2024–2025, that ratio was 18 to 1. OpenAI scrapes 1,500 times per referral. Anthropic 60,000 times per referral. AI is taking content and sending back a tiny fraction of the visitors. If you don't manage to get inside the "recommendation circle" before the door closes, there's a real chance you'll end up locked out.

What this actually means for your business

Here comes the good news. Preparing for AI search isn't a brand-new discipline — it's a smart evolution of what you already do (or should be doing). Danny Sullivan, Google's official Search Liaison, summed it up on the Search Off the Record podcast (December 2025): "I see AEO, GEO and similar acronyms as a subset of SEO, much like local SEO."

Czech SEO consultant Pavel Ungr added on his blog (December 2025): "In the last few months we've been buried under new acronyms: AIO, GEO, AEO. You don't need to invent 'AI SEO.' You need to do good SEO that also works in the age of AI."

Priority pyramid (1 = highest ROI)

  1. E-E-A-T authority — a named author with a photo, bio, contact details, and verifiable credentials. AI evaluates who stands behind the content.
  2. FAQ sections + FAQ schema on every key page. Format: question in an H3, answer in 40–60 words.
  3. A TL;DR summary of 60–90 words placed right after the H1. Per Pew Research, the median length of an AI summary is 67 words — copy that format.
  4. Structured data — LocalBusiness, Organization, FAQPage, HowTo, Review. AI is an extraction engine — make its job easy.
  5. Concrete numbers in your content. The Princeton GEO Paper (2024, an analysis of 10,000 queries) showed that adding statistics raises the odds of being cited in an AI answer by 30–40%. Keyword stuffing, by contrast, lowers AI visibility.
  6. External citations from authoritative sources — the Czech Statistical Office, government ministries, .edu, .gov, professional chambers.
  7. Activity on third-party platforms — Reddit, Quora, Heuréka, Diskuze.cz, industry forums. 85% of AI brand mentions come from these sources, not from your website.
  8. Wikipedia (if your company meets Wiki's "notability" criteria) — the single most-cited source in AI Overviews, period.
  9. Google and Heuréka reviews — the aggregate sentiment shapes how AI characterizes your business.
  10. English versions of your key pages. LLMs are trained primarily on English-language data. The way AI Overviews currently work in the Czech Republic is that Gemini translates the Czech query into English, generates the answer, and translates it back — so English-language sources have a leg up even on Czech queries.

Czech-market specifics

The Czech ecosystem has its own playing field:

  • Marketing Miner (founded by Filip Podstavec) is the only Czech tool that measures AI Overviews visibility through residential Czech proxies. Tomáš Novák of Marketing Miner put it this way: "Real visibility today means becoming part of the answer itself."
  • Most-affected Czech industries: health 32% of queries with AI Overviews, finance 24%, automotive 14%.
  • Most-cited Czech sources in AI Overviews: DrMax.cz, Pilulka.cz and Alza.cz (the leading Czech e-commerce players in pharmacy and electronics), cs.wikipedia.org, and Wikiskripta.eu. If you can find a PR angle or expert-byline route onto one of those, it's a high-leverage move.
  • The overlap between Google's organic top 10 and AI Overviews in the Czech Republic is only 17% (Shoptet Blog). Ranking #1 on Google no longer guarantees that you're in the AI answer.
  • Seznam.cz (the Czech-built search engine) is testing its own AI (a 70B-parameter model trained on Czech data), with launch planned for Q1 2026.

A balanced view — don't fall for the hype

The market is currently flooded with new acronyms: AEO, GEO, AIO, LLMO, SEE. Consultants are promising "a revolution" and "a brand-new SEO." The truth is more measured — and more useful.

AI search is a new entry point, but it's not the end of traditional search. Datos × SparkToro's State of Search 2025 Q2 showed that in the US and EU, Google still holds roughly 95% of traditional search. AI tools account for 0.64% in the US and 0.78% in the EU. AI traffic today represents roughly 1% of total web traffic (Conductor 2026).

But it's growing at +527% year over year (Previsible AI Traffic Report, January–May 2025). And 165× faster than organic (Forbes Agency Council 2026). And more important than today's volume is when the door closes. Winning a "brand mention" inside the training data of an LLM is cheap right now. In a year or two it won't be. And in five years, traditional search, in the words of OpenAI CFO Sarah Friar, will be "blowing open" — unsustainable in its current form.

For a business owner in their 40s or 50s, this is good news. You don't need to tear up your strategy overnight. You just need to do good SEO with the awareness that today, the reader is also an AI. And start now — because the window in which you can buy your way into AI recommendations cheaply is narrowing fast.


What's next?

Preparing your business for the era of AI search isn't about buying expensive tools or chasing acronyms. It's about getting content, authority, and brand right — well enough that AI has a reason to cite you, and the trust people give it works in your favor.

If you'd like to know how your business looks today through the eyes of ChatGPT, Perplexity, and Google AI Overviews, WebZaDen runs a free AI visibility audit. We manually test 30 queries relevant to your business, compare you against your competitors, and hand you a concrete list — what to fix now and what to work on for the long term. Estimate the cost of your project, or just drop us a line — we'll get back to you within 24 hours.