Claude's B2B Referral Share Jumped From 1.4% to 18.5%

Key Takeaways

• Claude’s share of measurable B2B AI referral traffic rose from 1.4% to 18.5% across two 2026 measurement waves, a roughly 13x increase that pushed it past Gemini and Perplexity into second place behind ChatGPT.
• A separate, broader dataset (SE Ranking) puts Claude at just 1.4% of all AI-referred traffic, but that’s measuring a different, more consumer-mixed denominator, not a contradiction.
• ChatGPT still holds 62.6% of measurable B2B AI referrals, so this is additive to your tracking, not a replacement for ChatGPT optimization.
• Claude’s audience skews technical, so specifications, honest comparisons, and docs-style structure tend to outperform marketing copy there.
• Bottom line: Claude has crossed the threshold from noise to a genuine, growing B2B channel, so start measuring your own citation rate on it directly, for example with Am I Cited , rather than assuming your ChatGPT visibility carries over.

The Number That Should Change Your Tracking List

For most of 2025, if you were building an AI visibility program, the platform priority list was obvious: ChatGPT first, everything else a distant second. New 2026 data says that list needs an update.

According to Goodie’s 2026 AI Search Traffic Report, Claude’s share of measurable B2B AI referral traffic rose from 1.4% to 18.5% between two measurement waves eight months apart, a roughly 13x increase that pushed Claude past Gemini and Perplexity into the #2 spot, behind only ChatGPT. Over the same window, ChatGPT’s own share of B2B AI referrals fell from 89.1% to 62.6%.

That is not a rounding-error move. It’s the kind of shift that, if it showed up in organic search, would trigger an emergency SEO review. The question is whether it deserves the same urgency in your AI visibility tracking, and the honest answer requires looking past the headline number.

Ready to Monitor Your AI Visibility?

Track how AI chatbots mention your brand across ChatGPT, Perplexity, and other platforms.

Get Started Contact

What Actually Changed, By the Numbers

Goodie’s report is built on two measurement waves: Wave 1 covered May–August 2025 (2.8 million AI referral sessions across 41 brand sites), and Wave 2 covered March–April 2026, using GA4 referrer data from a fresh anonymized B2B brand panel, triangulated against SimilarWeb’s platform-level traffic (25.77 billion visits, January–April 2026) and SensorTower’s iOS rankings.

Every major platform except ChatGPT gained share, but the gap between Claude’s growth and everyone else’s is the real story:

• ChatGPT: 89.1% → 62.6% (down 26.5 points)
• Claude: 1.4% → 18.5% (up 17.1 points, a ~13.2x multiple)
• Gemini: 2.4% → 10.6% (up 8.2 points, a ~4.4x multiple)
• Perplexity: 3.1% → 7.3% (up 4.2 points, a ~2.4x multiple)
• Copilot: 3.2% → 4.0% (up 0.8 points, a ~1.25x multiple)

The “Big 4” platforms together still account for roughly 99% of measurable B2B AI referrals, so this isn’t a story about a new entrant. It’s a story about redistribution among platforms you already track. And the redistribution isn’t noise: Goodie’s statistical testing found 14 of the 16 tracked brands moved in the same direction on Claude share, with a one-tailed p-value of 0.21%, which is a strong signal the shift is real rather than a sampling artifact.

Why the Number Looks So Different Elsewhere

If you go looking for a second opinion on Claude’s growth, you’ll find one, and it will look nothing like 18.5%. SE Ranking’s 16-month Google Analytics analysis, published in June 2026, puts Claude at just 1.4% of all AI-referred traffic, behind ChatGPT (78.23%), Perplexity (9.33%), Gemini (6.85%), and Copilot (3.57%).

Read quickly, that looks like a contradiction. It isn’t.

The two studies are measuring different denominators. SE Ranking’s figure is Claude’s share of all AI-referred site traffic across a broad, consumer-mixed panel, and Claude is used far more for writing, coding, and analysis than for the kind of casual, click-through search that generates referral traffic, which caps its raw click volume. Goodie’s figure is Claude’s share among the four major AI platforms, on sites that sell to businesses: the exact audience Claude’s product already over-indexes on.

Both numbers are correct. The practical takeaway is this: Claude is still a rounding error in total web clicks, but it is no longer a rounding error in B2B buyer research specifically, and B2B buyer research is the audience most B2B SaaS teams actually care about.

Stay Updated on AI Visibility Trends

Get the latest insights on AI mentions, brand monitoring, and optimization strategies.

Email address

Subscribe

Why B2B Specifically, and Why Now

Claude’s growth isn’t evenly distributed across use cases, and that unevenness is the point. A few contributing factors line up with the data:

1. Claude’s audience already skews technical. Developers, analysts, and technical evaluators are heavier Claude users than the general population, and those are precisely the people running vendor comparisons, reading documentation, and researching tools for their employer: B2B research, by definition.
2. A visibility spike had a real trigger. SE Ranking traces part of the March 2026 surge to public attention following Anthropic’s statements on Pentagon-related restrictions, a reminder that AI platform visibility can move on news events, not just gradual product improvement, which is part of why AI search predictions for the category keep getting revised mid-year.
3. Claude’s answer format rewards depth over breadth. Long, single synthesized answers with no ranked list favor content that thoroughly resolves a technical question, which is a different bar than optimizing for a ten-blue-links results page.

None of this means Claude is about to overtake ChatGPT in raw volume. It means Claude crossed a threshold, from “not worth tracking separately” to “a channel with its own measurable, growing share,” for exactly the buyer segment most B2B and SaaS brands are trying to reach.

What This Means for Your GEO Program

If your generative engine optimization strategy has been ChatGPT-first with Claude as an afterthought, this data is the trigger to rebalance, not to abandon ChatGPT, but to stop treating Claude as noise.

1. Track Claude as its own platform, not a ChatGPT proxy. Claude runs its own independent retrieval and doesn’t lean on Bing or Google’s index the way some assistants do. A strong AI share of voice on ChatGPT tells you nothing about your citation rate on Claude: you have to measure it directly. We go deeper on the methodology, including why single-run checks are misleading for a platform that produces probabilistic, synthesized answers, in our guide to tracking brand visibility in Claude .

2. Write for the evaluator, not the browser. Claude’s traffic skews toward people doing technical due diligence. Specifications, honest tradeoffs, and documentation-style structure tend to outperform persuasive marketing copy, a pattern we’ve detailed in our guide to Claude AI optimization .

3. Structure content to be citable in one answer. Claude gives a single synthesized response, not a ranked list. Self-contained sections, clear comparison tables, and clearly named sources give the model something concrete to cite, rather than something to summarize loosely.

4. Re-check monthly, not quarterly. A platform whose share moved 13x in eight months can reshuffle your competitive standing just as quickly in the other direction. Baking a Claude-specific check into your regular AI visibility monitoring cadence is cheap insurance against finding out three months too late.

The Bottom Line

Claude didn’t overtake ChatGPT, and it isn’t about to. What changed is that Claude stopped being statistically indistinguishable from zero in B2B AI referral traffic, and for a platform tilted this heavily toward technical, professional use, that shift lands squarely on the audience most B2B and SaaS brands are optimizing for in the first place.

The brands that will benefit aren’t the ones with the flashiest ChatGPT presence. They’re the ones who read numbers like these and immediately went to check whether Claude mentions them at all, because right now, most brands genuinely don’t know. Finding out is the first step, and it’s exactly what Am I Cited is built to measure.