The answer replaced the list
Google, ChatGPT, and Perplexity answer the question on the results page. Two or three brands get named. About two-thirds of searches now end without a click.
Search changed. Revenue didn't.
Buyers look in Google, in AI Overviews, in ChatGPT. We measure who controls demand in your category and build visibility where it pays. Reported in traffic, leads, and revenue.
How search changed
Six things we learned from reading server logs, running monitoring experiments, and following the latest research. Most sites are still built for the old version of search.
Google, ChatGPT, and Perplexity answer the question on the results page. Two or three brands get named. About two-thirds of searches now end without a click.
A brand already in the model's answer is cited about five times more often than one that is not. Being known beats being findable.
How often a brand is talked about across the web predicts AI visibility far better than its link profile does.
In six months of logs, AI assistants fetched pages many times more often to answer a live question than to train. Every fetch is a buyer asking.
AI referrals arrive on the homepage or site search, not on the page that was cited. The cited page did its job before the visit.
When Google demotes a section, the AI citations go with it. There is no separate AI visibility to optimize in isolation.
What we do
Every service moves a named component of the Demand Control Index and is measured against the same competitors every month.
How we work
Most agencies work like this: you pay, you wait, you get a report. With us you can see what is happening at every stage: the starting point, the priorities, what shipped, what it did.
Experiment protocol
We go through your site, your competitors, and what Google, AI Overviews, ChatGPT, and Perplexity show about the brand. The start is measured in numbers, not impressions.
A list of moves ranked by impact against effort. A shared dashboard from day one, not in the next report.
Technical fixes, content, AI Search work. Something ships every week, not once a quarter.
Every month we measure traffic, citations and revenue against the same competitors, and adjust what is not working. No charts for the sake of charts.
We take what works and extend it to more queries, more pages and more markets.
Lab notes
Nobody outside the companies that build the models knows exactly how they pick their sources, and what works changes fast. The only way to keep up is to keep testing, keep measuring and keep reading. That is the lab's day job, and what holds up goes straight into our clients' work.
Station 01
We run our own tests on how Google and AI assistants pick and cite sources. Every prompt runs several times, because one run proves nothing.
Station 02
Server logs, rankings and citations, measured over time instead of guessed. That is how we saw AI assistants fetch pages many times more often to answer a live question than to train.
Station 03
We read the new papers, patents and platform changes, test them, and turn what holds up into practice.
Who we are
We are experts in our fields and we know what we are doing. We have years of experience in large and small companies across different industries, from international marketplaces and classifieds to SaaS, in dozens of markets. Now we put that experience to work for our clients.
SEO & technical SEO
Content Strategy
AI Search & builder
Contact
30 minutes about your search and AI visibility. If it helps, we can prepare a short look at your site and competitors for the meeting.
Pick a time or leave your details below, and we will set up the call. In thirty minutes we talk through where buyers find you, who they find instead, and what we would do first. You leave with a clear read whether you work with us or not.
Prefer to pick a time yourself? Book a call
Or email us: hello@trafficlabs.agency
Or tell us where to look