Most of the argument about AI search is about loss. Will ChatGPT and Google’s AI Overviews strip out your clicks, gut your traffic, and end SEO as we know it. That’s the wrong thing to worry about first. The more useful question is what happens to the visitor who does land on your site from an AI answer. Early data says that person is the most pre-qualified, highest-converting visitor you’ll get all week, and most Australian websites are built to waste them.
Content Overview
The AI visitor arrives already sold
When someone clicks through from ChatGPT, Perplexity, or Claude, they’ve already done the part your website usually has to do. The AI has answered their question, weighed up a few options, and pointed them at you. They arrive at the end of the research, not the start of it.
The click itself tells you how much has already happened. Someone asks an AI for the best commercial electrician in Perth, or which project management tool suits a small team. The model reads across dozens of sources, filters out the noise, and hands back a short list with you on it. By the time they tap your link, the comparison is done and you have been recommended. That is a warmer introduction than almost any other channel gives you, paid or organic.
The conversion data is early, but it points one way. In a 2025 Seer Interactive case study of a single client’s site, visitors from ChatGPT converted at 15.9%, against 1.76% for Google organic. That is roughly nine times the rate, from the same site, over the same seven months. Perplexity came in at 10.5% and Claude at 5%.
One client isn’t a market, so the bigger sample matters more, and it tells the same story. Adobe Analytics, which tracks more than a trillion visits to US retail sites, found AI-referred shoppers converted 42% better than other traffic in March 2026. A year earlier the same visitors converted 38% worse, so this is a recent flip, not a settled law. Those AI shoppers also stuck around, with 12% higher engagement, a 27% lower bounce rate, and 48% longer on the site.
It’s a trickle, and that’s the point
The volume is tiny, and that matters. In the Seer study, AI sources drove about 11,000 sessions against nearly 14 million from organic, which is roughly 0.07% of the total. You are not replacing Google with this any time soon.
The trend underneath it is steep. A July 2026 study of 6.77 million AI-driven sessions across 166 sites found monthly volume had grown nearly tenfold in about a year and a half, with ChatGPT driving 92% of it. Seer estimates AI could reach 5 to 10% of organic traffic within a year for businesses that move early.
So this is a small stream of your most valuable visitors, and it gets bigger every month. Judging it on volume today misses what it is. You judge a channel like this on what each visitor is worth, and these ones are worth a lot. In the Seer study those 11,000 sessions still drove 1,370 conversions. At an average deal worth $150, that is over $200,000 of pipeline from a channel most businesses aren’t even tracking. Scale that to your own numbers before you decide it’s too small to bother with.
Your site is built for the wrong visitor
You’ve earned the click. Now the pre-sold buyer lands on a page built for someone who has never heard of you.
Most websites are designed to convince a cold Google browser. A long hero section, a brand story, a slow build of features and benefits, a bit of education, then eventually an ask. That is the right shape for top-of-funnel traffic that needs warming up. It is the wrong shape for someone the AI has already warmed up for you. They came ready to act, and the page asks them to start deciding all over again.
Picture a Perth services business whose homepage opens with a big brand statement, three scroll-lengths of “why choose us”, and a contact form tucked in the footer. For a stranger off a Google search, that is a reasonable journey. For a buyer an AI just told to enquire, it is a wall. They wanted the next step and got a company brochure. Most of them won’t scroll to find it, and they’ll go back to the chat and pick the option whose site made it easy.
There is a structural version of this problem too. That same July 2026 study found about a quarter of AI-referred visitors land on a site’s internal search page rather than a proper landing page. The model trusts your domain but can’t always pick the right URL, so it drops people at your search box and lets them navigate. On most sites internal search is an afterthought, and when it is, that high-intent visit turns into a bounce.
How to build a landing experience that closes a pre-sold visitor
The fix isn’t a rebuild. It is tuning the landing experience so it meets a decided buyer instead of stalling one. The same goes for AI agents: most sites don’t need a rebuild for those either. This is where generative engine optimisation hands off to user experience and conversion work, and four things do most of the lifting.
Match the page to the question. Send AI-cited content to a page that answers that specific query and offers the next step, not your homepage. If an AI keeps citing your pricing page or a comparison article, that page is now a landing page, so treat it like one: clear, current, and pointed at an action. Make your internal search genuinely useful too, because a real share of this traffic will use it to find their way.
Put your proof where they land. The AI vouched for you, and the page has to back that up in the first screen. Reviews, real results, client names, credentials, and an obvious sign there are humans behind the business. Australian buyers in particular want to see you’re local and real before they hand over an enquiry, so don’t make them dig for it.
Be fast. A pre-sold visitor on a slow page is a lost sale. If your site takes more than three seconds to load on a phone, you are leaking the exact people you worked hardest to earn. Core Web Vitals and mobile performance aren’t technical housekeeping here, they are conversion.
Give them one clear next step. The decision is made, so don’t bury the action under five competing buttons. One next move: book, buy, enquire, or call. Make it the loudest thing on the page, keep the form short, and make it work on a thumb. Every extra field and every extra choice is a chance for a ready buyer to change their mind.
None of that is exotic. It is disciplined web design aimed at a buyer instead of a browser, and it happens to make the page better for every other visitor too. If you want a hand working out which of your pages an AI is actually sending people to, that is the kind of job we take on.
Should you act on this now?
Not every business needs to move on this today. The honest read depends on what you sell.
If you’re already investing in getting cited by AI, through generative engine optimisation, and you’re not checking what those visitors do when they arrive, you’re paying to fill a leaky bucket. Fix the landing experience first. If you sell considered, higher-value services, or you run an online store where people research before they buy, this traffic is disproportionately yours. That is exactly the kind of question people take to an AI.
If you’re a low-consideration, impulse-buy business, or you can’t yet see any AI sessions in your analytics, this is a watch-and-prepare job rather than a drop-everything one. Set up the tracking, get a baseline, and revisit in a quarter. That last point is worth pausing on, because when we went looking across our own client base, most businesses can’t see this traffic properly yet, and the standard analytics setup hides a lot of it.
The evidence is still forming, and it’s worth being clear about that. The standout conversion numbers come from small or single-site samples. The big retail sample only flipped positive in the last year. Nobody has cleanly measured conversion rate by platform at scale yet. What isn’t in doubt is the direction and the logic. A visitor the AI has already convinced is worth more than one you have to convince from scratch, and building for that visitor is low-risk, because a clearer, faster, more trustworthy page helps everyone who lands on it.
The click is only half the job
Getting mentioned by AI is half a strategy. The GEO work gets you into the answer. The web and UX work turns that click into a customer. They are one job, and treating them as two is how your best visitor slips through the gap.
So before you spend another dollar chasing AI visibility, open your own site the way one of these visitors would. Land on the page an AI would send them to, and ask a plain question: is this built for someone still deciding, or someone ready to buy?
If you want a hand joining the GEO and the conversion sides, let’s talk.
Tom Porter
Studio Lead
