On Google you compete with nine other tabs. In ChatGPT there are no tabs. The AI picks one name, hands it over, and the search is finished. The signals it uses to pick that name aren't the ones you've been told mattered for Google.
Updated March 15, 2026
Around two thirds of Google searches now show an AI Overview at the top of the page. Industry trackers report that the majority of business-related searches touch an AI tool of some kind, whether that's a direct ChatGPT prompt or a Google result with the Overview attached. Semrush and others have published projections that LLM-driven traffic will rival conventional Google search within a few years.
None of that is a forecast about a distant future. It's a description of what's already happening this quarter.
When someone asks ChatGPT to recommend a plumber, or asks Perplexity for a branding consultant in their city, the answer doesn't arrive as a list. It arrives as a suggestion. And if your business isn't recognisable to those systems, you don't end up in the suggestion.
The interesting shift isn't really technical. It's about what the user sees.
A Google results page is a list. Ten links. The user scans, compares, opens a few tabs, and chooses. You're competing against nine other entries, which is uncomfortable, but at least everyone gets a chance to be one of the ten.
An AI answer is a recommendation. Sometimes one name. Sometimes a short list of three. There is no page two. Either you're in the answer or you aren't there at all, and the user gets on with their evening without ever learning you existed.
That's the whole game in a sentence: Google lets the user evaluate ten candidates; AI evaluates the candidates for them. The signals that won you a place on the old rankings still help, but they aren't enough on their own. The AI is looking for the kind of structure, credibility, and consistency a salesperson would look for before recommending someone to a client they like.
Four things matter most.
The first is structured data. The Schema.org markup that sits invisibly inside a web page and tells machines what they're looking at. Person, LocalBusiness, Service, Review, FAQ. Pages cited by AI tools have schema markup at meaningfully higher rates than pages that don't get cited; the technology is built to lean on it, because it removes the guesswork.
The second is content the AI can actually read. Not a list of links and a profile photo, but real text that describes what you do, who you do it for, and what you've delivered. AI tools need something to summarise. A wall of buttons isn't summarisable.
The third is domain authority. Content on your own domain pulls more weight than the same content sitting on a shared platform. Sites with even modest organic traffic are dramatically more discoverable to AI tools than pages with none, and the threshold is much lower than SEO professionals make it sound.
The fourth is consistency. Your name, location, services, and contact details should match across your site, your Google Business profile, your social profiles, and any directories you appear in. AI tools cross-reference. Conflicting information makes them cautious about recommending you.
Most of the popular link-in-bio tools provide none of those signals.
Linktree, Beacon, Carrd, and the rest generate close to zero structured data. They're well-designed for organising links, which is what they were built for, but they don't tell an AI tool what kind of professional you are or what evidence backs that up. To a crawler, a link-in-bio page looks like a hallway with twelve unlabelled doors.
The SEO benefit also accrues to the platform, not to you. Every time someone visits linktr.ee/yourname, the visit quietly compounds into Linktree's authority. None of it builds yours. You're a tenant on someone else's domain, paying rent in trust signals.
So when a prospective client asks ChatGPT to recommend someone in your field, the system genuinely can't tell you exist. Not because the work isn't good. Because the version of you it can read is twelve links and a profile photo, and that doesn't summarise into a recommendation.
Some businesses are starting to see real referral traffic from ChatGPT, where six months ago there was effectively none. Google's own data shows small and medium businesses with structured data picking up meaningfully more visibility in local results than those without. The numbers vary by industry and source. The direction does not.
What's worth paying attention to is the stickiness. Once an AI tool learns to recommend a particular professional in a particular field, that pattern tends to harden, in much the same way that first-page Google rankings became harder to displace over time. The estimates floating around put the comfortable first-mover window at roughly twelve to eighteen months. The number is soft. The direction is not.
The practical implication is small but worth saying out loud. The structured data needs to exist somewhere, on something you own, before the window closes. NoTrouble generates it automatically for every profile, which is partly why people use it. A developer can write the JSON-LD by hand, which works just as well. The tool isn't really the interesting question. The interesting question is whose domain the signal lives on, because the platform you put it on is the one that ends up getting the credit for it.
If your work depends on reputation and referrals, the action list is short and unglamorous.
Land on your own domain. The custom domain is the hook every other signal attaches to. Without it, every trust signal you build is being filed under someone else's name.
Put actual content on the page, in plain text. Describe what you do and who you help. Cleverness reads to a crawler like an unfinished thought; clarity reads like a fact.
Get your reviews visible. AI tools treat aggregated reviews as a primary trust signal, and they're one of the easier signals to wire up. Stale content, on the other hand, reads to a crawler as a business that may no longer exist.
Be specific about what you do. The AI is matching a query to a candidate. "Coach" loses to "executive coach for first-time founders" every time, because the second one is easier to map onto a person who just typed a sentence into a chatbot.
None of this is glamorous. None of it has to take long. The professionals who get this right in the next year or two will quietly become the names AI tools keep returning to. The ones who don't will still rank somewhere on page two of a Google that fewer people are bothering to scroll.
A professional presence with auto-syncing reviews, AI discoverability, and 36 design combinations — ready in 5 minutes.
Google gives you a list and lets you decide. AI tools give you an answer. Either you're in the answer or you aren't, and the signals these systems use lean heavily on structured data, reviews, and clear content rather than just keywords and links.
It's a small block of machine-readable data on a web page that describes the page's contents in a vocabulary search engines and AI tools agree on. It tells them you're a Person or a LocalBusiness, lists your services and reviews, and removes the guesswork. NoTrouble generates it automatically; most other tools require you to write it yourself.
No. Recommendation patterns are still forming, and the rough consensus is that there's a twelve to eighteen month window before the picture sets. The sooner there's solid structured data on your own domain, the sooner AI tools start treating you as a known entity.
No, but somebody does. The schema markup, the domain configuration, and the structured content all need to exist; the only question is whether you write them by hand, hire a developer, or pick a platform that produces them automatically. NoTrouble is in the third category. So is Wix's newer AI visibility tooling. Linktree, Beacon, and most of the link-in-bio category aren't in any of them.