Jerry.ai
US car-insurance super app
Using Sitefire, Jerry’s AI referral traffic rose 78%.
Jerry wanted to know where AI models were sending insurance shoppers, and what it would take to get Jerry’s own pages into those answers. Sitefire mapped the topics worth competing for, then showed page by page what was keeping Jerry out of them.
Jerry’s team shipped the changes in two waves between April and June 2026. The recommendations helped the team produce content with clearer structure, more direct answers, and better-sourced data on pages that were already live.
Every measure rose as a result. And because Jerry left a comparable set of pages untouched, there’s a clean read on how much of the lift was Sitefire’s insights and how much was the market moving on its own. The gap between them is the Sitefire effect.
+78%
AI referral traffic
Clicks to jerry.ai from AI answers.
+56%
AI bot traffic
Requests from AI models fetching Jerry's pages.
+27%
AI citation share
Jerry's citations as a share of all AI citations.
+9%
AI visibility
Share of AI answers mentioning Jerry.
Each figure compares the three months after the first improved pages went live with the three weeks before.
Singling out the Sitefire effect
In early March, a GPT-5.4 release cut AI bot traffic across the web, and this also affected Jerry. To recover traffic, Jerry used Sitefire to improve a selected set of pages, while keeping the rest untouched. The results is shown below:
Weekly AI bot traffic, March to June
Each group shown against its own level before the March update.
Following the changes in April, the improved pages recovered quickly from the GPT model update, drew up AI traffic and referrals, and ended up at an average of 112% of their pre-GPT model update level, compared to only 72% for untouched pages.
Sitefire tip: AI models re-crawl on their own schedule, so the numbers don’t move the day you publish. Expect a lag of a few days to a few weeks before the changes show up.
Why measure AI bot traffic and referrals?
AI visibility is the score you optimize against, but it’s calculated from a synthetic prompt set that someone had to choose. That choice is a hypothesis about what your customers ask, and a hypothesis can be wrong.
Traffic, on the other hand, is the ground truth. When referrals and bot activity move in line with visibility, the prompt set is tracking what real people are actually asking. When they don’t, the problem is the prompts, not the pages.
Jerry is spearheading the new era of agentic marketing
Using Sitefire, Jerry runs a continuous loop: find the prompts its customers ask AI, see where the gaps are, then fix those pages and publish.
Diagnose.
Sitefire researches which prompts are worth focusing on, then shows which brands the answers mention and which pages the models cite.
Improve.
For each topic, Sitefire agents research what’s working, then return one prioritized brief per page with line-by-line improvements. Jerry’s writers review the changes and publish.
Measure.
Sitefire tracks Jerry’s position on those prompts, and how AI bot and referral traffic respond to the edited pages.
Using Sitefire agents, Jerry can run this loop multiple times a day. Sitefire produces the briefs; Jerry writes and edits the content.
“Sitefire showed us which pages actually mattered for GEO and what each one was missing. The goal was never to publish more - we already have a pretty strong content production engine. Instead, we wanted to make the pages we already had get trusted and cited by the LLMs.”
Ida Sultan
Chief of Staff, Jerry
Jerry’s AI visibility and citation share have risen every month since March
Sitefire measures visibility and citation share against Jerry’s prompt set. The charts run March through July, so they cover the weeks before the first improved pages went live as well as the months after.
AI visibility
Citation share
The bottom line
Using Sitefire, Jerry turned their articles into content AI models cite consistently. Sitefire shows which of your pages AI models cite, what each one is missing, and what changes once you fix it. Using Sitefire agents, you can scale this approach to your whole sitemap.
Methodology
- The headline figures. Each compares the three months after the first improved pages went live with the three weeks before. Those three weeks fall after the March update, so the comparison chart indexes every week to the pre-update level instead.
- The comparison. The pages Jerry improved against comparable pages on the same site that Jerry did not touch, over the same weeks. The two groups did not fall equally in March, so each is indexed to its own level before the update and read against that level rather than against the other group. The chart runs March to June.
- AI bot traffic counts AI models fetching a page to build an answer to a live prompt. AI referral traffic is a floor, not a ceiling: many buyers research inside the model and then navigate directly, so they never appear as referral traffic at all.
- AI visibility and citation share are measured against a set of prompts fixed at onboarding and never changed, so later comparisons stay meaningful. Definitions at sitefire.ai/docs/kpis.
Get started with Sitefire
Take action to win customers in AI Search
