Harden the site you already have.
Still Fox, a Seattle practice, offers hardening, which fixes what keeps AI assistants out of the website a small business already has.
Hardening keeps the site the business has and fixes what stops AI assistants from reading it. In five business days the crawlers get in, the text sits in the HTML, the name, address and phone match everywhere, the missing service pages exist and the accessibility pass is done. A before-and-after page shows each change.
What changes
- Crawler access fixed in robots.txt and at the CDN or firewall for OAI-SearchBot, PerplexityBot, Claude-SearchBot, Meta-WebIndexer, Googlebot and Bingbot, verified by fetching with each user agent.
- Pages whose text appears only after JavaScript get a server-rendered version or a static mirror.
- Name, address, phone and hours made identical across the site, with a gap list for the directories we do not control.
- Missing service and location pages added, each opening with who, what and where.
- A WCAG 2.1 AA pass on the affected pages, automated plus manual, with a defect list for the rest.
Crawler access comes first because it is the fastest fix. A site that blocks OAI-SearchBot is out of ChatGPT search answers, and OpenAI says a robots.txt change can take about 24 hours to apply.
Hardening works on any platform where robots.txt, the CDN or firewall rules and the service pages can be edited: WordPress, Squarespace, Webflow and most agency-built sites. A CDN block is a settings change. Text that exists only after JavaScript runs needs a server-rendered version of the affected pages, or a static mirror of them.
Not included: a redesign, theme changes, plugin cleanup, hosting migration, content rewrites beyond the new service pages, and fixes to listings on third-party directories.
How it works
- Snapshot. We read the site first: robots.txt per crawler, the status each crawler's user agent gets from the CDN (a 403 or a challenge page is a block), the text with and without JavaScript, the name, address and phone, and a WCAG quick pass. The result is a ranked fix list. If it finds nothing to fix, we sell nothing.
- Baseline and scope. Before any fix, we ask a fixed set of buyer questions on the official search APIs of three assistants, several times each. That is the before. The fix list becomes the scope, at a flat fee.
- Fix. We apply the fix list in five business days. Each item ends marked done or blocked, with the reason.
- Re-check. We fetch every page again with each crawler's user agent and ask the same questions.
- Hand over. The client receives the fix list, the before-and-after snapshot, the directory gap list and the accessibility defect list.
How it ends
The client keeps the site, the hosting and every account. We change settings and pages inside them and write down each change, and the fix list records what was done and what is blocked, with the reason.
For a law firm, the lawyer signs off on every page we add or change before it goes live. Washington RPC 7.1 governs lawyer advertising, and the lawyer stays responsible under RPC 5.3 and 8.4(a).
When the platform cannot serve text to a crawler, or the business cannot get admin access to its own site, we say so and recommend a rebuild.
How it is proven
Crawler access is a pass or fail check. We fetch every page with each crawler's user agent before and after, and a 403 or a challenge page counts as a block. Rendering is a comparison of the text in the raw HTML with the text in the rendered page. Both are reported item by item.
The question panel runs on the official search APIs of OpenAI, Perplexity and Anthropic, with each question asked several times. It reports the share of answers that name the business with a 95% interval. A hardening can remove every block and still move that share by less than the interval, and the page reports it either way. We do not measure the consumer apps for clients.
For agencies
Hardening fits the clients an agency already has: they keep their site, and the agency keeps them. Still Fox does the work wholesale and white label. The agency sets its own retail price, Still Fox invoices the agency, and the before-and-after page is the agency's to hand over.
Fee
A flat fee per site, for the fix list the snapshot produces. Work outside the list is a change order, also at a flat fee. Measurement is included and never billed on its own.
Questions
What if the snapshot finds nothing wrong?
Then we sell nothing. The snapshot is free, and we report a clean result as it is.
Will you change my design or move my hosting?
No. Hardening keeps your theme, your platform and your host. It changes robots.txt, CDN or firewall settings, the pages that depend on JavaScript, and the facts that disagree across the site.
My robots.txt looks fine. How can crawlers still be blocked?
A CDN or firewall can refuse a crawler's request whatever robots.txt says, and robots.txt never shows it. We fetch the home page with each crawler's user agent and read the status code; a 403 or a challenge page is a block. Then we fix it at the layer where it happens.
Does it work on WordPress, Squarespace or Webflow?
Yes, on any platform where robots.txt, the CDN or firewall rules and the service pages can be edited, which covers most agency-built sites. If the platform cannot serve text to a crawler, we say so and recommend a rebuild.
Will clearing the blocks get my business recommended?
It removes a reason not to be named, and it is the part we can verify directly. A site that blocks OAI-SearchBot is out of ChatGPT search answers, so clearing that block is a precondition for appearing there. Whether an assistant then names the business is the assistant's call, and we report the share of answers with an interval, not a promise.
Specification and rates
Get a reading of what the assistants say about your business.
The snapshot is free and arrives by email in two business days.