4 min read
Dossier Site Foundation
How the Wacky Works redesign put clear pages, search signals, and reusable page patterns underneath the visual style.
About this piece
- Stage
- Studio note
- Published
- 2026-05-25
- Updated
- 2026-06-11
- Tags
- AI-ready websites, Structured data, Migration
Clear pages before decoration
Search signals matched to the content
Launch checks before public release
The work behind the look
The redesign started with the parts visitors rarely notice until they go wrong: clear pages, sensible redirects, search-friendly descriptions, feeds, and a reusable page pattern. The visual direction came after the structure was ready to carry real traffic.
That order mattered. A website can feel finished because the homepage looks good, but search systems, AI retrieval tools, social previews, and returning visitors all experience more than the first screen. They need stable routes, clear page purpose, useful metadata, sensible internal links, and public files that tell crawlers what the site contains.
The foundation work was therefore less glamorous than the visual system, but it was the part that made the site safe to grow.
Why it matters
An AI-ready website is not only a good-looking page. Search engines, AI answer systems, and human buyers all need the same basic thing: a site that explains what exists, where to go next, and why each page deserves to be there.
Putting those decisions in one shared foundation keeps the site easier to grow. New service pages, product pages, and articles can inherit the right behaviour instead of becoming one-off pages that drift away from the rest of the site.
For a small agency site, that shared foundation does three useful jobs. It reduces launch risk, because new pages use the same route and metadata rules. It improves discoverability, because sitemap, feed, AI discovery, and schema outputs come from the same source of truth. It also keeps the writing honest, because the machine-readable layer has to match the page a visitor can inspect.

The page map
The first decision was the page map. The site needed to explain capabilities, proof, practical writing, lab work, paid consultations, and the project intake route without turning every surface into another homepage.
Each route has a job:
- Capability pages explain what the studio can build.
- Field Reports show real studio work without inventing borrowed proof.
- Dispatches answer practical AI, automation, and search questions.
- Consultations sell a focused paid working call.
- Transmission handles project briefs and Test Pilot interest.
- Legal and product support pages stay plain and specific.
That page map gives humans a clearer navigation path, but it also gives search systems a better set of signals. A page can only rank or be cited properly when it has a reason to exist.
The search layer
The search layer was built as part of the foundation rather than sprinkled on at the end. Titles, descriptions, canonical URLs, structured data, social image routes, feeds, sitemap entries, image sitemap entries, robots rules, redirects, and AI discovery files all need to agree.
The rule was simple: the public machine-readable files should describe the live site, not the internal project plan. That meant removing development language, private implementation details, and vague claims from places that crawlers can read.
For AI-search visibility, the site also exposes llms.txt, llms-full.txt,
and .well-known discovery files. These do not replace classic SEO. They give
AI systems a cleaner summary of the most important public pages, while the
sitemap and feeds continue to serve traditional crawlers and readers.
The reusable page pattern
The page pattern keeps repeatable parts consistent: hero structure, capability summaries, evidence cards, FAQ blocks, related links, schema, and footer navigation. That does not mean every page looks identical. It means every page has the minimum furniture needed to be useful.
This matters during a launch because rushed pages tend to drift. A new article can accidentally miss a date. A service page can forget related links. A product page can make a schema claim that is not visible in the copy. A shared pattern reduces that drift.
What shipped
What shipped
- A clear page map for visitors, redirects, navigation, and search discovery.
- Page descriptions and structured data that match what visitors can actually inspect.
- Public discovery surfaces for search engines, AI assistants, social previews, and feed readers.
The lesson
The visible design carries the feeling of the site, but the foundation carries the trust. If a small business wants a website to be understood by people, Google, Bing, and AI answer systems, the work has to include the boring public interfaces too.
That is why a useful redesign should ask structural questions early. What pages deserve to exist? What should old URLs do? Which pages should be indexable? Which pages are private flow plumbing? What facts should schema expose? What content is strong enough to ask search engines to crawl?
Answering those questions before decoration gives the visual system somewhere solid to land.
Related capability
AI Search & SEO-Ready Websites
Websites built for people, search engines, AI assistants, and conversion from the same foundation.
