We Tracked AI Search Referrals for 28 Days. Here's Which Engines Send Real Traffic.
AI search engines are sending real traffic to websites. But how much? From which engines? And is it actually worth optimizing for?
We tracked every AI search referral to our B2B blog over the past 28 days. Here's the raw data — our baseline measurement for what we plan to track over the coming months as this channel evolves.
The Current State: AI Sends 1.6% of Our Traffic
Five AI search engines are sending measurable traffic to our blog (188 published posts on inblog.ai/blog):
| AI Search Engine | Visits (28 days) | Share of AI Traffic | Industry Share* |
|---|---|---|---|
| ChatGPT | 37 | 41% | 78% |
| Claude.ai | 18 | 20% | Not widely tracked |
| Gemini | 15 | 16% | 6.4% |
| Brave Search | 12 | 13% | Not widely tracked |
| Perplexity | 9 | 10% | 15.1% |
| Total AI | 91 | 100% |
*Industry share data from SE Ranking AI Traffic Research Study (2025-2026) and MediaPost (April 2026).
For context: 91 AI visits represent 1.6% of our total 5,518 visits during this period. The industry average is approximately 1.08% according to SE Ranking. We're slightly above average — likely because our content is heavy on CMS comparisons and SEO guides that AI engines frequently reference.
How Our Distribution Differs from Industry Data
The industry data says ChatGPT dominates with ~78% of all AI referrals. On our blog, it's only 41%. The difference comes from two sources most reports don't even track:
- Claude.ai at 20% — This is significant. Claude appears to be especially active in citing technical and comparison content. Most AI referral studies don't break out Claude separately.
- Brave Search at 13% — Brave's AI-powered search sends more traffic to us than Perplexity. This is a blind spot in most analytics setups.
Meanwhile, Gemini's 16% share far exceeds its 6.4% industry average on our blog. As of April 2026, Gemini has overtaken Perplexity as the #2 AI referral source globally (MediaPost, April 3, 2026). Our data appears to reflect this shift early.
The Conversion Rate Argument
The volume story (91 visits) sounds unimpressive. The conversion story changes everything.
According to Metricus (2026), AI referral visitors convert at dramatically different rates than traditional search:
| Referral Source | Avg Conversion Rate | vs. Google |
|---|---|---|
| Claude.ai | 16.8% | 6.0x |
| ChatGPT | 14.2% | 5.1x |
| Perplexity | 12.4% | 4.4x |
| Gemini | ~3% | 1.1x |
| Google (traditional) | 2.8% | baseline |
Source: Metricus, ChatGPT Referral Conversion Rates 2026
Why the massive difference? AI search users tend to arrive with higher intent. When ChatGPT or Claude recommends a specific tool or resource, the user has already been through a conversational research process. They're not browsing — they're following a recommendation from a source they trust.
Applying these conversion rates to our traffic: our 91 AI visits may generate roughly the same number of conversions as 450+ traditional Google visits. That reframes AI search from "rounding error" to "emerging high-value channel."
What Gets Cited by AI Engines?
Based on our analytics, certain content types appear more frequently in AI referrals:
- Comparison and alternatives posts — Our "Ghost Alternatives" and "Medium Alternatives" posts appear to be heavily cited. AI engines love structured comparison data.
- How-to guides with specific numbers — "How many internal links per page" gets AI referrals because it answers a specific, factual question.
- Data-backed case studies — The Ahrefs growth strategy post provides the kind of concrete data AI engines prefer to cite.
The common thread: content with clear, structured, citable data points. AI engines prefer content that makes definitive statements backed by evidence over general advice articles.
Comparing AI Referrals to Other "Small" Channels
To put AI referrals in perspective, here's how they compare to other traffic sources on our blog:
| Source | Visits (28 days) | Share |
|---|---|---|
| AI Engines (total) | 91 | 1.6% |
| Bing | 17 | 0.3% |
| Naver | 7 | 0.1% |
| DuckDuckGo | 5 | 0.1% |
| YouTube | 4 | 0.1% |
AI search already sends 5x more traffic than Bing to our blog. And unlike Bing, AI search traffic is growing rapidly month over month. If current growth rates hold (~1% MoM according to SE Ranking), AI referrals could reach 3-5% of total traffic within a year.
What We're Doing About It
Based on this baseline data, here's our optimization plan:
- Monthly tracking. We're committing to recording AI referral data monthly to track the growth curve. This post is our baseline — expect updates.
- Structured data in every post. Tables, numbered lists, and clear definitions make content more citable by AI engines.
- FAQ sections. Adding FAQ schema helps both traditional SEO and AI citation.
- Direct, definitive statements. AI engines prefer "The best X for Y is Z because..." over "It depends on many factors..."
Key Takeaways
- AI search is real. Five engines are sending measurable traffic to our blog today. It's 1.6% of total — small but growing.
- The distribution is more diverse than you think. ChatGPT leads, but Claude, Gemini, and even Brave Search are significant. Track them separately.
- Conversion rates change the math. AI referral visitors convert at 5x the rate of Google organic. 91 visits ≈ 450 Google visits in value.
- Comparison and data-rich content gets cited most. If you want AI engines to recommend your content, give them structured, citable data.
- This is the baseline. We'll be tracking this monthly and sharing updates as the channel evolves.
For the broader context on how AI search is changing the CMS landscape, read our State of Blog CMS in 2026 report.
Data source: inblog Analytics referrer tracking, March 16 – April 13, 2026. Industry benchmarks cited from SE Ranking and Metricus (2026).