Peec AI
Google Search ConsoleThe client
The client is a health-tech platform that helps consumers judge the quality of products they buy online, using data it collects and publishes itself. It is careful with its data and chose not to go public with this work at all. So this study carries no name, and every result is a multiplier or a percentage, measured in the client's Search Console, the AI prompt panel we ran for it, and our weekly reports.
When the work began, the product already had a loyal following in online communities and a growing brand search. Search beyond its own name was not yet a channel the business could rely on.
The problem
The platform had data nobody else had, and a search presence that did not match it.
- Most search traffic was its own name. People who already knew the brand found it; people searching the category mostly did not.
- There was no content program. No editorial plan, and no pages built around the questions buyers were asking.
- AI engines rarely put it first. For the safety and quality questions it exists to answer, AI answers named better-known names ahead of it.
- The data was not reaching the people who cite it. Raw results sat in the product, where neither journalists nor AI systems picked them up.
The strategy
We ran a dual-track program: every page written to rank on Google and to be quoted by AI engines. One rule held for all of it: the content had to earn trust, because trust is what the platform sells. Volume was never the goal.
- Articles built on the client's own data. Our writer worked directly with the founder, and every claim, figure and sentence was checked by the founder before it went live.
- Category pages. Hub pages that answer the category's core searches, built around what the test data shows.
- Comparison pages. Pages for people searching for other brands in the market, so the platform's independent data appears beside the brand they asked about.
- An entity-first structure for AI engines. Structured data, question-led sections and an llms.txt file that states plainly what the company is and does.
- Weekly measurement. Search Console, rank tracking and the AI panel reviewed every week, with the next pages chosen from what moved and what dropped.
The results
Google impressions. Daily impressions in June 2026 were 3.9× the daily average of the month before the work began. The steepest climb came from March, as the new pages were indexed, and it kept rising through June.
Google clicks. Daily clicks grew 2.7× over the same window. March alone brought more than twice the clicks of the month before the work began.
Rankings. Keywords ranking in Google's top three grew 10×, and keywords on page one 5.4×, from the month before the work to June (Ahrefs).
Estimated organic traffic. Ahrefs' traffic estimate, which models visits from rankings, grew about 6× from December to June.
Comparison searches. The site now ranks in the top five for several other brand names in its market, reaching people who were searching for a competitor.
Sign-ups. Clicks from Google into the page where people start the free product grew 15× in a single month, from February to March.
AI answers. In early April the platform was the first brand cited on 17 of the 30 buyer prompts we tracked in Peec AI (57%). By late May it was the second most-cited brand in its category, with 11% visibility, a 24% share of voice (up 6.1 points in 30 days) and an average position of 2.2 when named. Semrush measured the audience of the topics where AI engines mention it growing 6.6× from February to March.
Press. Three major news outlets cited the platform's original data in their reporting, with no outreach and no paid placement.
The honest caveat
Search was already growing before the work began: daily impressions doubled from October to December 2025, and the chart shows those months so the change of pace is visible. The chart's April, May and June come from weekly and monthly totals rather than single days. The Ahrefs figures are estimates, not counts. The AI readings are single windows from Peec AI and Semrush.
What made this different
This was not a traffic program with a health topic. It was a trust program that happened to rank. The press citations came because an article carried original data that nobody else had published, not because of its metadata. The AI track was built into the content from the first week rather than added afterwards, which is why the platform is named in AI answers and not only in Google results.
