Your buyers used to type your brand name into Google. Now they ask ChatGPT about it. Here's what that shift breaks and what you should be tracking instead.
Last updated: June 4, 2026
Key takeaways
- Branded search volume has been declining since 2023 for a meaningful share of B2B and consumer categories. Buyers increasingly ask AI engines about brands instead of navigating to brand websites directly.
- This breaks one of the longest-running proxies for brand awareness: monthly search volume for your brand name in Google Search Console.
- AI-mediated brand awareness is harder to measure but no less real. A buyer who asks ChatGPT about your brand and reads a positive description is forming an impression you can't see in any traditional analytics tool.
- The new metric isn't branded search volume. It's brand mention rate across AI engines for the prompts your buyers actually type.
- Trademark owners and brand operators should adjust their measurement stack accordingly. Branded search is still useful but no longer sufficient.
What used to happen and what happens now?
Until roughly 2023, the standard buyer journey for a known brand went something like this. The buyer became aware of the brand through marketing, word of mouth, or a referral. When they were ready to evaluate or buy, they typed the brand name into Google. Google returned the brand's website at position one, plus a few ads, a few news results, and a knowledge panel. The buyer clicked the website and the visit was attributed in analytics.
That sequence still happens, but it's competing with a new one. The buyer becomes aware of the brand, and instead of typing it into Google, they ask ChatGPT (or Perplexity, or Gemini) a question about it. "Is Brand X any good?" "How does Brand X compare to Brand Y?" "What does Brand X actually do?" The AI returns a synthesised answer. The buyer reads it. They form an impression. They make a decision. They never visit your site.
Your analytics don't record this. Google Search Console shows your branded search volume slowly declining. You don't see the AI conversation that replaced it. That's the death of brand navigational search in practical terms: the search query has been replaced by a buyer-AI exchange you can't see.
How much has branded search actually declined?
The numbers are real but noisy. Several independent analyses have been tracking this:
- Search Engine Journal reported in 2025 that B2B SaaS branded search queries had declined 10 to 20 percent year over year for many tracked accounts, with the steepest drops in technical product categories.
- Search Engine Watch published similar findings on consumer brands, with branded discovery searches ("how does Brand X work") declining faster than direct navigational searches ("Brand X login").
- SEMrush's 2025 industry report flagged a measurable shift from informational and navigational queries to transactional and bottom-funnel queries on Google, consistent with informational journeys moving to AI engines.
- Ahrefs published research showing that ChatGPT's training cutoff and retrieval behaviour mean newer brand information disproportionately gets surfaced through prompts rather than direct navigation.
None of this means branded search is going to zero tomorrow. It does mean the long-standing assumption that branded search volume is a clean proxy for brand awareness is breaking down.
Why does this matter for trademark owners?
Trademarks exist to give buyers a way to recognise your brand and distinguish it from competitors. USPTO defines a trademark as something that "identifies the source of the goods and distinguishes them from others". The recognition mechanism the trademark protects has historically run through the buyer's eyes and ears: they see your logo, they hear your name, they recognise it.
That recognition mechanism now also runs through AI engines. A buyer might never see your logo. They might never hear your name in an ad. They might form their entire impression of your brand from a ChatGPT answer that summarises what AI thinks of you. The trademark you spent years building is being filtered through a synthesis layer you don't control.
For a brand operator managing a portfolio of registered marks (see our owner profiles for examples of brand portfolios with hundreds of filings each), the implication is straightforward: brand investment alone no longer guarantees brand recognition in 2026. The AI engines need to know your brand and need to surface it positively when buyers ask.
What replaces branded search as a measurement?
Three honest candidates, listed roughly in order of usefulness:
1. AI mention rate for category-defining prompts. For each major category your brand competes in, identify 5 to 15 prompts your buyers actually type. Track whether ChatGPT, Perplexity, and Gemini mention your brand in the answer. This is the cleanest signal that the AI layer "knows you" the way your buyers expect.
2. AI mention rate for branded prompts. "Is Brand X legitimate", "Brand X reviews", "what does Brand X cost". These are the navigational queries that have moved from Google to AI. Track whether the AI answer is broadly accurate and positive. This is the cleanest signal that the AI layer "describes you correctly".
3. AI citation share. When buyers ask about your category, which URLs does the AI engine cite to support the answer? Are they URLs you own, sites where you've earned coverage, or completely unrelated sources? This is the cleanest signal for "where should I invest in earned and owned coverage to influence the AI".
All three need a tool that captures real AI engine output. They can't be measured by traditional brand monitoring, social listening, or SEO tools.
How is this different from social listening?
Social listening tracks what humans say about your brand on public platforms. It's a useful signal for sentiment, news velocity, and crisis detection. Brandwatch, Meltwater, Cision, and others all do this well.
AI monitoring tracks what AI engines say about your brand to the humans who ask. The relationship between the two is meaningful but indirect. AI engines absorb content from public sources, including social platforms, and synthesise it into answers. So a wave of negative tweets about your brand will eventually colour the AI engine's tone, but the connection isn't immediate or symmetrical.
The practical difference: social listening tells you about the public conversation. AI monitoring tells you about the private buyer conversation that the public conversation eventually shapes. For brand operators in 2026, you want both, but if you can only afford one, AI monitoring is closer to revenue.
Does any traditional metric still matter?
Yes, several. Direct traffic to your website is still meaningful. Branded search volume, even if declining, is still useful as a trend signal. Social mentions, news mentions, and earned media coverage all still feed into the AI layer.
The change isn't that these metrics stop working. It's that they no longer give a complete view. A brand can have stable branded search volume and falling AI mention rate at the same time, and the second number is the leading indicator of the first one continuing to drop.
A practical rule for 2026: if your existing brand measurement stack doesn't include any AI engine output, you've got a gap. Adding even 10 to 20 weekly AI prompt checks fills it for most categories.
What should brand operators actually do?
Three concrete moves:
- Add AI mention rate to your brand KPI deck. Even if you only track 10 prompts per major brand, having the number on the dashboard gets it the attention it deserves. Use a tool with real chatgpt.com data rather than API data, because the two diverge.
- Map your filing strategy to your AEO strategy. The trademark classes you file in shape what AI engines associate with your brand. Use our Compare Strategies tool to see how your filing pattern compares to your competitors' and identify where the AI engine might be confusing the two.
- Invest in coverage on the surfaces AI engines pull from. Reddit, LinkedIn, YouTube, and trade press disproportionately get cited for bottom-funnel buyer questions. These are durable surfaces, less ephemeral than chasing weekly news cycles, and they compound.
The brands that adapted early to the search engine era kept their brand awareness through the transition. The brands that adapt early to the AI engine era will do the same.
Frequently asked questions
Is branded search going to disappear completely?
Unlikely in the near term. Buyers will still use Google for many tasks where the answer involves navigation (logging in, downloading apps, locating addresses). What's declining is the discovery and evaluation portion of branded search, which is moving to AI engines.
How do I tell whether my branded search decline is due to AI or just market changes?
Compare your branded search trend against a category benchmark over the same time period. If your branded search is down 15 percent year over year but the category overall is down 12 percent, the gap is small and probably market-driven. If your decline is 25 percent and the category is 5 percent, you have a brand-specific issue that AI monitoring may help diagnose.
Can I influence what ChatGPT says about my brand?
Indirectly. AI engines pull from the public web. Investing in coverage on the surfaces they cite most heavily for buyer-decision queries (Reddit, LinkedIn, YouTube, trade press, owned blog content with question-format headings and FAQ schema) does shift what they surface over time. It's slower than running an ad campaign but more durable.
What's the relationship between AI visibility and SEO rankings?
Weak. HubSpot's 2026 research found backlink count has near-zero correlation with AI citations. Page rank for a query and AI citation for that query are two different problems. Some content does both well, but optimising for one doesn't reliably optimise for the other.
Should I worry about AI engines describing my brand incorrectly?
Yes, if you sell something where description matters (pricing, capability, regulated industries). The fix is the same as for press misinformation: publish accurate, structured, citable content on your own site and in trade press, and monitor whether AI output reflects that content over time.
How does this connect to trademark filing strategy?
If AI engines start describing your brand using terms that overlap with competitor brands, that's both a marketing signal and a potential dilution signal. The defensive response is to file additional protective marks where appropriate and to publish content that reinforces your distinctive positioning. Track filing patterns of competitors in the same category via the Compare Strategies tool or relevant NICE class pages.
Want to know whether AI engines actually surface your brand the way your buyers expect? Join the AEO Tracker waitlist. We onboard a few brands at a time and report weekly on real chatgpt.com data.