
A website can look great and still be hard for search engines and AI assistants to read. When that happens, your business is less likely to show up when someone searches, and less likely to be quoted when someone asks an AI tool for a recommendation. Our site releases now ship with a standard set of search checks, verified on the live site, not just in a preview.
Key takeaways
- Search engines and AI assistants read the code behind your pages, not the design, so small technical gaps add up.
- Every page needs one clear main heading, a unique title and description, and a canonical tag that says which address is the real one.
- Social share cards control what people see when your link is posted or texted, and they help your pages look trustworthy.
- An llms.txt file gives AI tools a plain summary of who you are and what you offer.
- Checks only count when they are run on the live site after launch, because previews often differ from what the public sees.
Why can a nice-looking website still be hard to find?
Search engines do not see your photos and colors first. They read the page code: the title, the description, the headings, the links, and a few small tags most owners never look at. AI assistants that answer questions do the same. If those signals are missing or confusing, the tools have to guess, and they often guess in favor of a competitor whose site is easier to read.
None of this replaces having a clear offer and a way to book. Our post on why business websites do not book appointments covers that side. This post is about making sure the people looking can find you at all.
Which checks matter most on every page?
These are the basics we confirm on every release:
- One main heading per page. A single heading-one that says what the page is about. More than one confuses the outline.
- A unique title and description. Each page gets its own, written for a person reading search results.
- A canonical tag. A one-line note in the code that says "this is the official address of this page." It stops search engines from splitting credit between near-duplicate pages.
- Social share cards. Tags that set the image, title and summary shown when your link is shared.
- A sitemap. A list of every public page, so search engines do not have to find them one link at a time.
- FAQ markup where it fits. Structured questions and answers that search engines and AI tools can lift directly.
- An llms.txt file. A short plain-text summary for AI crawlers. We explained it in making your site AI-ready with llms.txt.
What does this look like on a real site?
For a retail client in November, we ran a search scan before launch and worked through what it found. The membership page got a full title, description, social card and canonical tag. Every published blog post received a canonical tag and a social card, and each page was trimmed to one main heading. Where two posts covered the same topic, the weaker one now points its canonical at the stronger post, so search engines see one clear answer instead of two half answers. We also updated the automated blog publishers so every new post gets a canonical tag from the start.
We fixed the posts one at a time and read every other field back before moving on, so nothing was republished or reposted by accident. The scan's blocking findings went down after the first round, and on 2026-11-08 every published post returned a working page with exactly one canonical tag, a social card and one main heading.
We do not promise rankings. Nobody honest can. What we can promise is that the site gives search engines and AI tools clean, consistent signals to work with.




