Dual-channel visibility: why your website now has to be built for two audiences
btlabs Core · Jul 23, 2026
For as long as the web has existed, websites have been built for a single reader: the human. Design, navigation, classic SEO — all of it geared toward getting one person onto the page and keeping them there.
That assumption no longer holds. A growing share of your future customers will never read your website. They ask ChatGPT, Perplexity or Google's AI Overview — and get a finished answerin which your business appears. Or doesn't.
This isn't a distant future, it's measurable. And it means: from now on, your website has to serve two audiences at once — the human and the machine. We call this dual-channel visibility.
The end of the pure “human web”
A growing share of Google searches today end without a single click to an external website. Back in 2024, around 60% of US Google searches already did — out of 1,000 searches, only about 360 reached the open web (SparkToro/Datos, 2024). By early 2026 that figure had risen to 68% — barely a third of searches still reached the open web at all (SparkToro/Similarweb, 2026).
Once an AI Overview appears, it gets worse: according to Seer Interactive, organic click-through rate drops by 61%. And already across all searches in general, around 60% end with no click to an external source at all (Bain & Company, 2025). The user has their answer — in the AI’s chat window. The classic click to your site never happens. This phenomenon is called zero-click search.
The uncomfortable consequence: if the AI doesn't mention you in its answer, you're simply invisible to a large part of your audience — no matter how beautiful your website is.
The two channels
The two channels are the human channel and the machine channel: one relies on aesthetics, brand identity and trust, the other on clear facts and a machine-readable structure for AI crawlers and agents. A future-proof website therefore runs on two tracks:
The human channel. Aesthetics, brand identity, trust, a fast and accessible experience — without an annoying cookie banner. People don't buy because the data is technically clean, but because they're convinced by a brand. This channel stays important.
The machine channel. A second, structured layer made of clear facts, unambiguous entities and lean, token-efficient text — built for AI Crawlers and Agents. Free of menus, layout and scripts, optimised for maximum semantic clarity.
The point is: what's optimised for people (images, animations, nested layouts) is often hard-to-read ballast for an AI — and vice versa. Cleanly serving both channels from the same data foundation is the real task.
What the research shows
Two studies prove it: clear, substantive content gets cited by AI more often and more accurately than marketing fluff.
Classic SEO is no longer enough. The landmark study “GEO: Generative Engine Optimization” (Aggarwal et al., Princeton/Georgia Tech, ACM KDD 2024) was the first to quantify how content has to be structured to be cited as a source by AI search engines . The result: the right optimization can increase visibility in AI answers by up to 40% . What works best are concrete statistics, embedded expert quotes and clean source citations — substance instead of marketing fluff.
AIs still often get sources wrong. A study by the Tow Center for Digital Journalism (Columbia, 2025) tested eight AI search tools with 1,600 queries. In over 60% of cases the source was cited incorrectly — and the systems preferred to answer confidently but wrongly rather than admit they didn't know. That's exactly why your information has to be available so unambiguously and machine-readable that an AI never even has to guess.
The message of both studies is the same: if you present your facts clearly, with evidence and well-structured, you get cited more often and more correctly.
The dilemma: visibility versus data sovereignty
The dilemma: to be visible in AI answers, you have to make your content open and machine-readable — but doing so risks losing control over whether that content also gets used to train AI models. With AI visibility, a new fundamental question lands on every website operator: do you want your content to be used for training AI models?
Through technical standards like the W3C TDM Reservation Protocol (based on Article 4 of the EU Copyright Directive, referenced by the EU AI Act) you can declare a reservation of rights in a machine-readable way. Important here: under the working group's interpretation, search and discoverability are considered not covered by the opt-out. In plain terms: a modern approach lets you control this granularly — “visible in AI search: yes, model training with my data: no”.
This control is not a detail. It decides whether you take part in the AI web without giving away your most valuable content.
The new moat: original data
The new moat is called original data. In a world where AI generates any amount of generic text on demand, the value of standard content falls toward zero. The only lasting value lies in what an AI cannot invent: your original data, your local expert knowledge, real case studies and verified human voices.
That's the good news for small, specialized businesses. Your concrete knowledge about your market, your region, your customers — cleanly structured and readable for human and machine — is exactly the moat that interchangeable AI text never crosses.
What this means for your business
In practice you need a platform that serves both channels from a single, maintained data foundation: a fast, privacy-friendly experience for people — and in parallel a clean, machine-readable layer for AI systems, including granular control over what the AI may read and what it may use for training.
That's exactly what we built btlabs Core for. It's the bridge to dual-channel visibility: your company data maintained centrally once, delivered to website, shop and app — and machine-readable to AI. In up to six languages, without a cookie banner, on your own EU server.
And because honest beats loud: no technology guarantees you a mention in every AI answer — no provider who promises that is serious. What a clean foundation does is significantly raise the probability and make sure that, when you are cited, you are cited correctly . The substance is yours to deliver. We make sure it lands — with human and machine.
Sources & studies
- GEO: Generative Engine Optimization — Aggarwal, P. et al. (Princeton, Georgia Tech, IIT Delhi), ACM SIGKDD 2024. arxiv.org/abs/2311.09735
- 2024 Zero-Click Search Study — Rand Fishkin, SparkToro / Datos, 2024. sparktoro.com
- In 2026, Less than One Third of Google Searches Still Send a Click — SparkToro / Similarweb, 2026. sparktoro.com
- AI Search Has a Citation Problem — Tow Center for Digital Journalism, Columbia University, 2025. cjr.org
- W3C TDM Reservation Protocol (TDMRep) — W3C Community Group Final Report, 2024. w3.org/community/tdmrep
- The /llms.txt proposal — Jeremy Howard, Answer.AI, 2024. llmstxt.org
Note: GEO, llms.txt and TDMRep are young, evolving standards. No major AI provider has yet formally committed to weighting them as a fixed citation signal — the benefit is plausible and growing, but not guaranteed.
Frequently asked questions.
Why is classic search engine optimisation alone no longer enough?
Because a growing share of your customers no longer search the classic way — they ask AI assistants, and you only appear there if your website is readable for AI systems too. Classic search still brings visitors, but it is now just one of two channels. Serve both — people in the browser and AI systems as a second audience — and you get found wherever your customer asks. This principle is called dual-channel visibility.
Has classic SEO become obsolete with AI search?
No. Google AI overviews mostly cite pages that already rank well classically. SEO remains the ticket in — AI optimisation (GEO) is the extension with citable answers, structure and visible authority.
Do AI answers mean I lose website visitors?
Overall, fewer clicks come through — there's no denying it. The AI answers many questions directly, without anyone opening your site at all. This mostly affects simple questions like opening hours or basic prices.
So what matters is no longer raw visitor count, but whether you show up and get recommended inside the AI answer itself. If you show up there, you still win inquiries — just not through a classic website click.
Do I need to rebuild my website for AI visibility?
Not necessarily. Structured data, clean multilingualism and performance can be retrofitted. But if the foundation is an old tower of plugins, a foundation that brings these things built in is often the faster and cheaper way.
