The case study.
In January 2026 we started building a consumer education platform for a branch of a national retail mortgage lender. Eight months later, AI assistants have gone from citing it 1,126 times a month to 9,126 — an 8.1x climb that is still steepening. On ten of its strongest questions, it is the majority source, with citation shares as high as 87.5%. Every number below is measured, sourced, and bot-excluded.
What was built
312 indexed URLs in eight months — 187 pages and 125 articles, published at a sustained pace of more than 9 URLs per week. A multi-state product page system. Interactive calculators. Structured data on every page. Secure application intake wired to the client’s own systems. All of it on infrastructure the owner controls: their domain, their hosting, their analytics.
All figures on this page are pulled from the named dashboards as of September 4, 2026 (Google Search Console reports: September 4, 2026), and can be reproduced live on request. Page last updated September 4, 2026.
What it earned
Before and after the takeover
This platform existed before we took it over in January 2026 — it had traffic, history, and a full prior year of lead data. That makes it an honest experiment: same domain, same owner, one variable changed. Here is what moved.
Why none of this is bot traffic
We voluntarily excluded 92% of raw site traffic as automated before reporting a single number. Search Console measures activity on Google’s results pages, not our server — bots scrape websites; they do not run Google searches. AI citations are counted inside Microsoft’s systems when assistants answer real questions. Three independent sources, each bot-proof for a different structural reason. That methodology ships with every client engagement.
The signal money cannot buy: Google and Microsoft independently identified the same page as an authority. Search placement can be purchased. Being the source an AI assistant relies on cannot.
