What we are finding
Practical answer engine optimisation, written to be useful on its own, whether or not you ever speak to us.
-
How to prove a structured data change actually did something
Take a baseline before you change anything, because you cannot reconstruct one afterwards. Then measure three things: whether the markup validates and resolves, whether AI crawlers are fetching the changed pages in your server logs, and whether you are cited across a fixed set of customer questions asked repeatedly. Ask each question at least ten times — model outputs vary, so a single before-and-after screenshot proves nothing.
-
Structured data adoption is rising. Here is how to measure it yourself
Do not take anyone's word for adoption rates, including ours. Fetch the homepage and one deep page from ten competitors, count the JSON-LD blocks on each, and check whether they resolve into one connected graph or several unconnected ones. In most sectors you will find that presence is now common and correctness is still rare. Presence is becoming table stakes; the shape of the graph is where the remaining advantage sits.
-
Why small technical changes move an AI visibility score so much
Because most of the technical layer is binary. A crawler is either allowed or blocked, a page either renders without JavaScript or it does not, structured data is either valid and connected or it is ignored. These are switches, not dials, so a single line in robots.txt or one corrected JSON-LD block can flip an entire category at once. Content quality is the slow gradient underneath; the technical layer is where the sudden movement lives.
-
Schema.org for tour operators: the markup that actually matters
For a tour operator, four schema.org types do most of the work: Organization or TravelAgency for the company entity, Product with Offer for each trip and its price, FAQPage for question-led content, and BreadcrumbList for structure. Add Review and AggregateRating only where the reviews genuinely exist on the page. Build them as one connected graph per page rather than separate blocks.
-
How to check whether ChatGPT can read your website
Request your robots.txt and look for rules naming GPTBot, ClaudeBot, PerplexityBot or Google-Extended. Then fetch a live page using each crawler's user agent and check the HTTP status, because CDN bot protection often blocks them at the edge with no robots.txt rule involved. Finally, disable JavaScript and reload a product page to see what a non-executing crawler actually receives.