Websites that search engines and AI assistants can actually read.
Trevor Stuckless builds websites for local businesses in Ontario through Blue Crane Works, his company in Hamilton. Each site is written from the owner's own facts, one page per service, town or customer question, and served as plain HTML with every word in the first response. Before any release, a script refuses to ship a duplicate title, a dead link or an accidental noindex.
What does a website from Trevor include?
A page for each thing a customer asks about, instead of one page trying to say everything. Services, the towns served, guides for the questions people type late at night, the story of the business and how to visit, each at its own address with its own title, description and structured data.
The business facts are stored once. Hours, phone number, address and service area feed every page and every structured-data block from a single file, so the site cannot contradict itself or drift away from the Google listing. Alongside the pages sit the files machines look for: a robots file that names each crawler, a sitemap whose dates are true, an llms.txt summary, a Markdown copy of every page and a proper error page.
Why build pages as plain HTML instead of a JavaScript app?
Because the systems that decide whether a business gets found read the HTML first, and not all of them run scripts. Bing's Webmaster Guidelines list hiding critical content behind client-side rendering as something to avoid, and warn that such content may not be indexed or used in its AI answers.
Order matters too. Google documents that Googlebot reads the first 2MB of a page and indexes only what it downloaded, so the words, the sources and the structured data sit near the top of each file rather than behind a bundle of code.
How do you keep a site from quietly breaking after launch?
With a deploy gate: a script that runs before every release and fails it on the mistakes that cost visibility without anyone noticing. A repeated title, a missing Markdown copy, a link to a page that no longer exists, broken structured data, a stray noindex, or an error page that loads scripts.
That last one comes straight from Bing, which asks that 404 pages stay static and free of service calls. The gate on this site runs dozens of checks of the same kind, including one that fails the build if any two pages read too much alike.
Which websites has he built, and what did they measure?
| Site | What was built | Measured | Source |
|---|---|---|---|
| Garage Doorman Grimsby, two domains | A residential and a commercial site moved off Wix with the company email intact, every old address redirected to its closest new page | Commercial pages open to Google went from 0 to 36; 3,018 Google impressions in the first 13 days | Case study |
| Amarettos Dundas | Department, brand and guide pages for a shop that had a single page, with larger type and higher contrast for an older clientele | Indexable pages went from 1 to 28; 12 of 12 crawlers receive the full homepage | Case study |
| Referral Flooring Burlington | A four-step estimate request, an office portal for the enquiries, and flooring answers drawn from the installation standard | Live since the 3 September 2026 cutover | referralflooring.com |
| Dodge Daddy Hamilton Region | A parts catalogue for three Ram 1500 generations built around the owner's own badge, with guides that cite their sources | 41 pages at launch in September 2026 | dodgedaddy.ca |
| This site | A single scene page grew into ten pages with Markdown copies, crawl files and a deploy gate | Audit warnings went from 25 to 3 on 28 September 2026 | How it was audited |
What does a website cost?
A managed website from Blue Crane Works is $0 down and $100 a month in Canadian dollars on a twelve-month term, with hosting, edits and a quarterly visibility update included (published terms). Booking, quote intake, a store or a customer portal is $50 a month each. For a site the owner already has, the AI Visibility Pass is the one-time alternative.
What else does Trevor build?
- AI and search visibilityProving, crawler by crawler, that Google, Bing, ChatGPT and the rest can read a site, then fixing what cannot.
- Business applicationsSoftware that runs the work itself: job-site records, parts and stock, paper archives.
- AI featuresAI that removes one specific chore: site walks to estimates, paper to documents, websites an assistant can use.
- Payments and billingStripe subscriptions, one-time passes, donations, and platform payouts to the people on a platform.
- Data pipelinesReading hundreds of public sources on a schedule and handing back one thing a person can act on.
- Scheduling and eventsCheck-ins, calendars, rosters and maps for gyms, a festival and a members community.
- Media and advertisingSix media titles and the ad server that sells and counts every placement directly.
Talk to Trevor about a website.
Describe what the business needs and what it is doing by hand today. You will get a straight answer on whether this is the right job for it, and what it would cost.