Made Verdacert findable in AI answers on a Domain Rating of 4 - AI assistants out-sent organic search 2.8× in October

The brief

Verdacert's buyers are immigrants, paralegals and law firms with a USCIS, court or university deadline - and a growing share of them start by asking an AI assistant where to get a certified translation. Each order ships a signed certification statement and a public verification URL backed by a cryptographically signed receipt, so the product is built to be checked by machines as well as people.

It launched in May 2026 on a domain with almost no authority. As of November 2026, Ahrefs rates it a Domain Rating of 4. The brief was to make Verdacert legible to AI assistants and AI agents from launch week, so it could be found in the answers people actually read, and to measure that on checkouts.

The results

• AI assistants sent 2.8× organic search's visits in October 2026 (1,811 vs. 638), and have out-sent search every week since September 7.
• 3,236 visits from AI assistants since launch, through November 6, 2026, with headless-scraper traffic excluded.
• AI-referred buyers convert far better. Since checkout tracking began on October 4, AI visitors to customer pages triggered 30 checkout events from 194 visits; organic search, 8 from 436 - about 8× the per-visit rate.
• AI became the recruiting channel. 2,515 visits to translator job pages came from AI assistants - 8.6× organic search - for reviewer roles in languages like Rohingya, Somali, Oromo and Dari.

Who the AI visitors are

Not every AI visit is a buyer, and the case study is more useful if we say so. Most of Verdacert's AI traffic lands on reviewer job pages: people looking for certified-translation work. That is real value for a service that needs native speakers of rare languages, but it isn't revenue. On customer pages, organic search still sends more visits than AI assistants. What AI sends instead is intent: people who asked a specific question - questions like how to get a birth certificate translated for USCIS, or what a certified Farsi translation costs - and arrive ready to order.

What we built for assistants and agents

• An llms.txt that answers the buying questions. Following the llmstxt.org spec, Verdacert's llms.txt carries every price tier, turnaround time, guarantee and source language, with Markdown twins of key pages, so an assistant quotes Verdacert accurately instead of guessing.
• Explicit AI-crawler access. robots.txt names and admits GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended and Applebot-Extended.
• Machine-readable pricing. Organization, ProfessionalService, Service and Offer schema with unit-price specifications, and every advertised price rendered from the live pricing engine - one source of truth for pages, snippets and assistants.
• Agent-ready from day ten. An MCP server with 8 tools (quote, submit, track, retrieve, refund, case packets and certificate verification), listed on the official MCP Registry, with a /.well-known/mcp.json manifest, a REST API and OpenAPI spec, and free sandbox keys - so an AI agent can order and verify a certified translation without a human in the browser.
• Multilingual programmatic pages. 30,666 indexable URLs across seven site languages - English, Arabic, Farsi, Urdu, Turkish, Spanish and Haitian Creole - covering document types, language-by-document pages and guides.
• A page for every role. Each reviewer opening has its own indexable job page with JobPosting schema, which is what turned AI assistants into Verdacert's recruiting channel.

What we learned

• Low authority didn't block AI visibility. On a Domain Rating of 4, AI assistants were sending 2.8× organic search's traffic within five months of launch.
• Clean data beats big data. A headless scraper was 92% of GA4 users in September-October 2026 until it was filtered out. Without that filter, every channel comparison here would have been wrong.
• Measure AI on the outcome you sell. Visit counts would have overstated buyer demand from AI; checkout events show where AI is actually worth the most.

The same playbook drives LandEarly, where AI assistants send 2.4× organic search's visits and convert 2.4× better, and Raven Health, cited by all six major assistants.

Methodology

Traffic and conversion figures are from Verdacert's Google Analytics 4 property, pulled November 7, 2026, using GA4's default "AI Assistant" and "Organic Search" channel groups, with sessions at a 1366×1366 screen resolution (a headless scraper) excluded. Customer pages exclude careers, reviewer sign-up, sign-in and account pages in every site language. Checkout comparisons start October 4, 2026, the first day key events were recorded; "checkout events" are begin_checkout and purchase key events. Some purchases are recorded server-side without a channel, so AI-attributed purchases are a floor. Domain Rating is from Ahrefs; page counts are from verdacert.com's live sitemaps; the MCP listing is on registry.modelcontextprotocol.io.