GA4 Can Now Track AI Traffic — But Are You Tracking What Your Competitors Are Doing With It?

But here’s where marketers need to pump the brakes. The update is largely a repackaging of data GA4 was already collecting. And it comes with real blind spots. As MarTech noted, the new AI Assistant channel only works when GA4 can detect a referrer — meaning traffic from copied links, mobile apps, or in-app browsers may still appear as Direct traffic if referral data gets stripped before the visit reaches your site. Google also hasn’t published a full list of supported AI referrers beyond ChatGPT, Gemini, and Claude, leaving coverage for platforms like Perplexity and Microsoft Copilot uncertain.

The deeper limitation, though, isn’t technical — it’s strategic. GA4 is a rearview mirror for your own site. It tells you what arrived. It cannot tell you what you’re missing, what your competitors are capturing, or which content is earning citations you never see because the click never happens. As Semrush states plainly: “GA4 shows you what traffic arrived from AI sources. It doesn’t tell you how your traffic compares to competitors, or which content is earning citations in the first place.”

That distinction is critical. Most marketers will open their GA4 dashboard, see the new AI Assistant channel populating (or not), and draw conclusions from that single data point. If AI referral traffic is growing, they’ll assume their strategy is working. If it’s flat, they might deprioritize AI optimization entirely. Both conclusions are dangerous without competitive context — because the number that matters isn’t just how much AI traffic you’re getting. It’s how much AI traffic you’re not getting that your competitors are.

This is the blind spot the GA4 update inadvertently creates: a measurable channel that feels complete but covers only half the picture. And in a landscape where AI-driven discovery is moving from theoretical to measurable, half the picture is exactly the kind of gap that competitors can exploit while you’re busy admiring your own dashboard.

The Gap Everyone Is Ignoring — AI Traffic Is a Competitive Intelligence Problem, Not Just an Analytics One

Here’s a thought experiment. Imagine you run a B2B SaaS company, and your GA4 dashboard shows 1,200 AI-referred sessions last month — a 40% increase from the quarter before. That feels like progress. But what if your top three competitors each received 5,000? Suddenly your “growth” is actually a widening gap disguised as a green arrow on a report. This is the fundamental problem with treating AI-referred traffic as a site analytics question instead of what it actually is: a market share question.

Knowing your own AI referral numbers without competitive context is like knowing your conversion rate but not your category’s average — directionally interesting but strategically useless. You can celebrate a 3% conversion rate all day until you discover the industry benchmark is 7% and you’re actually underperforming. The same logic applies here. GA4’s new AI Assistant channel, as MarTech reported, gives marketers a cleaner view into how AI assistants drive traffic, making it easier to compare AI referrals with organic search and measure how those visitors convert. That’s valuable. But it’s a single-player view of a multiplayer game.

The real strategic question isn’t “how much AI traffic did we get?” It’s “who is getting cited instead of us, and on which queries?” The signals to answer that question already exist — they’re just sitting in tools most performance marketers haven’t connected to their workflow. Semrush’s guide on measuring AI search visibility explains how prompt-level tracking lets you pinpoint which prompts your brand dropped out of, which ones you’re newly appearing in, and which competitors gained ground on a specific query. That’s not abstract competitive theory. That’s a concrete, query-by-query map of where your share of AI-driven discovery is growing or shrinking relative to the brands you compete against.

Think about what this means in practice. If a competitor starts appearing in ChatGPT’s responses for “best project management tools for remote teams” and you don’t, that’s not something your GA4 dashboard will ever surface. You won’t see a dip in AI referral sessions because you can’t lose traffic you never had. The absence of signal is itself the signal — and it’s invisible to anyone whose monitoring starts and ends with their own analytics property.

This creates a dangerous lag effect. By the time declining AI referrals show up in your own GA4 reports, the competitive dynamics that caused the decline — a competitor publishing more citable content, earning mentions from sources AI systems already trust, restructuring pages in formats that LLMs prefer — have been compounding for months. If you only monitor your own GA4 channel, you’ll know you’re losing market share approximately six months after it’s already happened.

The chain reaction is predictable. As Semrush’s research on AI visibility notes, when AI visibility grows, branded search tends to follow, and when branded search grows, conversions tend to follow. That chain works in reverse, too. Lose AI citations to a competitor today, and you’re not just losing chatbot clicks — you’re losing the downstream branded search volume and homepage conversions that would have followed. The compounding cost of ignorance isn’t theoretical; it’s measurable across multiple reporting cycles, if you’re tracking the right signals.

GA4 tells you that AI traffic arrived. Competitive intelligence tells you why it arrived for someone else and not you. One is a rearview mirror. The other is a windshield. Most teams are still driving by looking backward.

What Top-Performing Advertisers Are Already Doing Differently

While most marketers are still configuring their GA4 filters, a cohort of top-performing advertisers has already moved past the measurement phase and into active optimization — building playbooks specifically designed to capture and convert audiences arriving through AI referrals. The difference between these early movers and everyone else isn’t budget or team size. It’s a three-part operational shift that treats AI visibility as a strategic channel rather than a reporting curiosity.

The first move: auditing which URLs earn AI citations and reverse-engineering the patterns. AI-cited content looks structurally different from traditional SEO content. Pages that consistently appear in AI-generated answers tend to feature direct, definitional statements early in the copy, use clearly labeled data points, and organize information in formats that language models can easily parse — think comparison tables, numbered frameworks, and concise expert attributions. Top advertisers are running systematic audits to identify which of their existing pages already receive AI referral traffic, then cataloging the shared structural elements across those pages. The goal isn’t to rewrite every blog post — it’s to establish a template that increases the probability of future citations. Once that template exists, content teams can apply it to new pieces targeting high-intent queries, and performance marketers can align native ad creatives and push notification copy to match the language patterns that AI platforms are already surfacing.

The second move: using prompt-level tracking to find competitive gaps. This is where the intelligence layer gets genuinely powerful. Tools like Semrush’s Prompt Tracking allow marketers to monitor specific conversational queries — the actual questions buyers type into ChatGPT, Gemini, or Perplexity — and see exactly which brands appear in the responses. By tracking prompts at the individual query level, advertisers can pinpoint where competitors are earning citations they aren’t, identify which third-party domains AI systems trust as sources in their vertical, and prioritize outreach or content creation accordingly. This isn’t theoretical. When you can see that a competitor consistently appears in responses to “best project management tool for remote teams” while you don’t, that’s an actionable gap you can close with targeted content, strategic partnerships with the cited sources, or paid amplification designed to build the brand signals that feed back into AI training data.

The third move: building correlation models that justify budget reallocation. This is where smart advertisers separate themselves from the pack. Because AI visibility often influences purchasing decisions without generating a click — someone reads a ChatGPT recommendation, closes the app, and searches your brand name directly the next day — last-click attribution in GA4 will systematically undercount its impact. The emerging best practice, as Semrush’s reporting framework outlines, is to track three signals in parallel over time: AI referral sessions in GA4, branded search volume via Google Search Console, and conversion rates on homepage traffic where most branded visitors land. When AI visibility grows, branded search tends to follow, and when branded search grows, conversions tend to follow. Documenting that chain across multiple reporting cycles builds the evidence base stakeholders need to greenlight budget shifts.

This correlation approach matters enormously for performance marketers running native and push campaigns. If your AI citations are driving a measurable lift in branded search — which HubSpot’s analysis of the category supports, noting that AI-referred visitors already convert at 4.4 times the rate of traditional organic visitors — then your paid campaigns benefit from a compounding awareness effect you didn’t pay for directly. The smartest advertisers are treating AI visibility the way they once treated podcast sponsorships or influencer partnerships: as an upper-funnel awareness channel whose ROI shows up downstream in cheaper CPAs and higher conversion rates on branded terms. If you’re only reading last-click data in GA4, you’ll see the effect without ever understanding the cause — and your competitors who do understand it will keep pulling ahead.

The Crawler Access Audit Most Marketers Haven’t Done

Before you benchmark a single session, before you compare your AI referral numbers to a competitor’s, there’s a prerequisite so fundamental that skipping it renders everything else meaningless: you need to confirm that AI crawlers can actually reach your content.

Here’s the brutal irony playing out across thousands of marketing teams right now. They’re celebrating GA4’s new AI Assistant channel, tweaking dashboards, and scheduling weekly reports — while their own robots.txt file is actively blocking the very bots that would generate that traffic. If ChatGPT-User, OAI-SearchBot, Perplexity-User, or Claude-SearchBot hits a disallow directive before it ever parses your page, your content doesn’t get indexed by those models. It doesn’t get synthesized into answers. It doesn’t get cited. And your shiny new AI Assistant channel sits at a flat, immovable zero — not because your content isn’t good enough, but because you’ve locked the door and walked away.

This isn’t a hypothetical edge case. When major CMS platforms and hosting providers rolled out updates over the past year, many included default robots.txt rules that block known AI user agents. Some site owners added those blocks intentionally during early debates about AI training data, then forgot to revisit the decision once AI-referred traffic became a measurable — and increasingly high-converting — channel. The result is a self-inflicted blind spot that no amount of competitive intelligence can fix.

The audit itself takes minutes. Open your robots.txt file and search for user-agent directives targeting ChatGPT-User, GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot, Claude-Web, Bytespider, and Google-Extended. If any of these are followed by Disallow: /, those crawlers are locked out entirely. If you find partial blocks — specific directories or URL patterns — evaluate whether those restrictions still serve a business purpose or whether they’re relics of a policy decision that predates AI traffic tracking altogether. For most content-driven sites, the highest-ROI action available today is removing or narrowing those blocks to let AI systems index the pages you actually want cited.

But even after you’ve cleared the crawler access hurdle, your numbers still won’t tell the full story. As MarTech has reported, the new AI Assistant channel only works when GA4 can detect a referrer — and traffic from copied links, mobile apps, or in-app browsers may still show up as Direct traffic because referral data gets stripped before the visit reaches your site. That means sites that do allow crawlers are almost certainly undercounting their AI-referred sessions. The gap between what GA4 reports and what actually happens could be substantial, especially for brands whose audiences skew mobile or whose content is frequently shared through chat threads and messaging apps.

Layer on another complication: Google hasn’t published a complete list of supported AI referrers. Coverage is confirmed for ChatGPT, Gemini, and Claude, but as MarTech noted, there’s still uncertainty around whether platforms like Perplexity or Microsoft Copilot are fully recognized by the new channel grouping. Meanwhile, Ahrefs found that AI chatbot traffic — while still representing a small share of total web visits — converts at rates that can dwarf traditional organic search, making every uncounted session a missed signal about your most valuable visitors.

The action item is non-negotiable: audit your crawler access before you start benchmarking, or your baseline is built on incomplete data. No competitive analysis, no prompt tracking strategy, and no dashboard redesign matters if the foundation — letting AI bots see your site — hasn’t been laid first.

Building the Competitive AI Traffic Stack — A Framework for Performance Marketers

The three-part operational stack that separates early movers from everyone else isn’t theoretical — it’s a layered system you can build this week using tools most performance marketing teams already have access to. Think of it as three tiers, each one expanding your field of vision from what’s happening on your own site to what’s happening across your entire competitive landscape at the individual prompt level.

Tier 1: Your Own Data via GA4. This is the foundation. With GA4’s dedicated AI Assistant channel now automatically categorizing sessions from ChatGPT, Gemini, and Claude using the ai-assistant medium, you finally have a clean baseline for AI referral volume without regex hacks or custom channel groups. Set up a dedicated exploration report filtered to the AI Assistant channel group, then layer in your conversion events — purchases, form fills, demo requests — to measure not just volume but value. Track session quality metrics like engagement rate and pages per session against your organic and paid benchmarks. This tells you whether AI-referred visitors are browsers or buyers. But here’s the critical limitation: GA4 only captures sessions where the referrer data survives the click. Traffic from copied links, mobile apps, or in-app browsers often appears as Direct traffic because referral data gets stripped before the visit reaches your site. That means GA4 is necessary but structurally incomplete — which is exactly why you need the next two tiers.

Tier 2: Competitive Benchmarking Across AI Assistants. GA4 tells you about your traffic. It tells you nothing about the traffic your competitors are capturing from AI systems — or the citations they’re earning that you’re not. This tier requires dedicated AI visibility platforms that estimate share of voice across 20-plus AI assistants for your category’s most important queries. The goal is to understand your brand’s presence relative to competitors across every major AI system your buyers use. As HubSpot’s analysis of the category makes clear, a brand can appear in 90% of prompts on one platform and be completely absent from another, so multi-platform tracking isn’t optional — it’s the only way to get an accurate competitive picture. Build a monthly scorecard comparing your AI citation frequency against your top three to five competitors across ChatGPT, Gemini, Perplexity, and Copilot. When you see a competitor gaining ground, that’s your signal to investigate what content or authority signals are driving their visibility.