AI search tools like ChatGPT, Claude, and Perplexity are already changing how buyers discover and evaluate brands — pulling content from your site without sending a visit, recommending brands by name, and soon transacting on users’ behalf.
Apiary Digital Founder, Karen Amundson, recently sat down with Senior SEO Consultant, Chris Burdick, to discuss how the rules of search have changed. Watch the full interview or keep reading for more.
Traffic is no longer a clean proxy for marketing success, and the teams winning right now are shifting content investment, rebuilding their measurement stack, and getting technically ready for AI agents.
Key Takeaways
- SEO isn’t dead. Most traffic still comes from Google. Keep investing in it.
- AI is hollowing out top-of-funnel content. Shift budget toward commercial-intent content: comparison pages, versus pages, buyer’s guides.
- AI search has no stable rankings. “Am I being mentioned consistently across queries my buyers care about?” is the right question now.
- A 30% drop in organic traffic does not mean a 30% drop in revenue. You’re in a measurement transition period.
- Brands that build AI presence now get baked into the next generation of models. There’s a real first-mover advantage that compounds over time.
What do SEO, AEO, GEO, and AGENTIC actually mean?
The acronyms are flying around LinkedIn like confetti. Here’s what they actually mean and why you should care.
- SEO (Search Engine Optimization) is what you already know. The blue links on Google. Backlinks, on-page optimization, technical health. Still the single biggest source of traffic for most brands. Don’t stop doing this.
- AEO (Answer Engine Optimization) is about getting cited at the top of AI-generated answers — think Google’s AI Overviews, featured snippets, voice search responses. When the AI answers a user’s question and includes a link back to your site as a source, that’s a citation. AEO is the game of earning those citations.
- GEO (Generative Engine Optimization) is about getting recommended by ChatGPT, Claude, Perplexity, and friends. When a chatbot tells a user “I’d recommend Brand A, Brand B, or Brand C,” that’s a mention. GEO is the game of becoming one of those recommended brands.
Agentic refers to AI agents that don’t just answer questions — they take action. The difference between asking ChatGPT how to book a flight and telling an agent to actually book it. Big difference for marketers, because the agent isn’t going to “visit” your site the way a human does.
Term | What it means | The goal
- SEO | Traditional search optimization for Google’s blue links | Earn ranked positions in search results
- AEO | Optimization for AI-generated answers (AI Overviews, featured snippets, voice) | Earn the link back as a cited source
- GEO | Optimization for ChatGPT, Claude, Perplexity recommendations | Be the recommendation itself
- Agentic | AI agents that take action on a user’s behalf | Be technically ready for agents, not just humans
If you walk away with one mental model, make it this — AEO is about earning the link back. GEO is about being the recommendation itself. Both matter. They require different work.
Why The Agentic Web Will Hit Marketers Hard
In the next 12 to 24 months, expect to see AI agents transacting on behalf of users. Someone tells their agent “buy me a new pair of running shoes under $150 with good arch support,” and the agent goes and does it. No website visit. No add-to-cart click. No abandoned cart email opportunity.
The protocols to enable this (things like the Agentic Commerce Protocol) are still being built. If you’re on a major platform like Shopify, much of this will be handled for you automatically. But the brands that win in 2026 and 2027 will be the ones whose websites are technically ready for both human visitors and AI agents.
Your tech stack and your site structure need to assume three audiences now — humans, traditional search crawlers, and agents.
Why AI Search Has No Rankings (And What to Measure Instead)
AI search doesn’t have rankings the way Google does. When you Google a query, you and your colleague in the next Zoom box see roughly the same results. Stable. Trackable. That’s why we’ve built entire industries around keyword rank tracking.
Run the same prompt five times in ChatGPT and you’ll get five different answers. Run it on your account vs. someone else’s and the answers will diverge again, because chatbots learn your preferences over time. The technical term is generative variance — the tendency of AI systems to produce different outputs for the same input — and it’s not a bug. It’s the entire architecture.
- Keyword rank tracking, as we’ve known it, doesn’t translate to AI search.
- Searching your own brand on your own ChatGPT account and feeling great about the results is a trap. Your account already knows you love your brand.
- “Am I ranking?” is the wrong question. “Am I being mentioned and cited consistently across queries that matter to my buyers?” is the right one.
What Most Marketing Teams Are Getting Wrong Right Now
Teams are applying SEO playbooks to AI search and wondering why it’s not working. The biggest example is top-of-funnel content strategy.
For two decades, the SEO formula has been: publish a lot of top-of-funnel informational content, capture the awareness audience, nurture them down the funnel. It worked. Beautifully.
AI is eating that top of the funnel. When someone asks ChatGPT “what is X” or “how does X work,” they get a synthesized answer. They never visit the informational article you wrote to capture them. The top of the funnel as a traffic acquisition strategy is being hollowed out.
Where to Move Your Content Investment
Shift more of your content budget to middle- and bottom-of-funnel commercial content:
- Comparison pages
- “Versus” pages
- “Best of” lists
- Use case deep-dives
- Buyer’s guides
There are two reasons this works:
- It captures commercial-intent users right now. When an AI does pull content into a response, it’s increasingly pulling commercial content because that’s what users are asking AI to help them evaluate.
- It trains the next generation of models on your brand. The more high-quality commercial content you publish about your category, the more future LLMs learn to associate your brand with your category. This is the long game.
The earlier you get serious about this, the more you bake your brand into the models that haven’t been trained yet. There’s a real first-mover advantage that compounds over time. Brands that wait are going to pay a premium to catch up.
Why a Traffic Drop Doesn’t Mean a Revenue Drop
If your organic traffic is down 20-30%, here’s what’s actually happening: a meaningful portion of the value your site used to capture as visits is now being consumed without a visit. An AI bot crawls your page, lifts the information, and serves it to the user inside ChatGPT or Perplexity.
The user got your content. Your analytics didn’t get the visit.
A 30% drop in organic traffic does not mean a 30% drop in revenue. People are still buying. They’re just discovering you through channels you can’t measure with the dashboards you built in 2019.
You cannot directly attribute revenue to a mention right now. The tooling isn’t there yet. So you have to stop using traffic as a clean proxy for marketing success.
We’re in a measurement transition period, and the leaders who get ahead are the ones who expand what they measure rather than clinging to legacy KPIs.
What Should You Actually Be Measuring For in AI Search?
Citation Tracking via Google Analytics
You can set up custom tracking in GA to filter referrals coming through ChatGPT, Perplexity, Claude, and other AI tools. Set this up today if you haven’t. It only tracks human visits coming from AI, but it’s the foundation.
An AI Visibility Platform
Tools like Profound and Otterly run hundreds of prompts across the major LLMs, monitor whether your brand is being mentioned and cited, and track how often your competitors show up too. This is the closest thing to “rank tracking” that exists for AI search, and it’s the metric we’d recommend most CMOs start watching alongside traditional organic.
Brand Sentiment
This is the underrated one. With traditional SEO, sentiment didn’t matter much — a link was a link. With AI, how people talk about your brand online dramatically affects what LLMs say about you. A flood of negative Reddit threads can quietly become how ChatGPT describes your product.
Share of Mention vs. Competitors
Not just “am I being mentioned” but “what percentage of relevant queries mention me vs. my top three competitors.” This is the closest thing to share of voice we have in the AI era.
Two Measurement Traps To Avoid
#1. Don’t Be Your Own Test Subject
Searching your own brand on your own ChatGPT account and concluding “we’re showing up great!” is misleading. Self-personalization is making the chatbot tell you what it knows you want to hear. Use incognito mode, logged-out sessions, or — better — a proper visibility tracker that depersonalizes the testing.
#2. Don’t Only Measure What You Want to Be Found For
Marketers are biased. We pick the queries we hope we rank for. The bigger opportunity is usually in queries you didn’t think of — queries your buyers actually use, in language that isn’t on your messaging doc.
The best visibility tools generate prompt sets using AI, which surfaces what you could be found for, not just what you assume you should be.
You don’t know what you don’t know. Build a measurement process that surfaces blind spots, not one that confirms what you already believe.
What Should Marketing Leaders Do In The Next Two Quarters?
- Keep investing in traditional SEO. Most of your traffic still comes from Google. Don’t burn down what works.
- Layer in AEO and GEO strategies. Shift content investment toward commercial intent — comparisons, versus pages, best-of lists, buyer’s guides. Stop relying on TOF content to do the work it used to do.
- Get your technical foundation agent-ready. Talk to your dev team about structured data, schema, and emerging agentic commerce protocols. If you’re on a major CMS, you’re probably most of the way there. If you’re on a custom build, you have homework.
- Build a real AI search measurement stack. Custom GA tracking. A visibility tracker. Brand sentiment monitoring. Share of mention against competitors.
- Reset expectations internally. Your CEO and board are looking at organic traffic the same way they did in 2022. Educate them on why that number is going to keep moving in ways that don’t track linearly with revenue, and build the new metrics into your reporting before they ask.
- Move now. The brands that establish presence in AI answers and recommendations today are the ones being baked into the next generation of models. This window won’t last.
FREQUENTLY ASKED QUESTIONS
Is SEO dead now that AI search is taking over?
No. SEO is still the single biggest source of traffic for most brands. Most of your traffic still comes from Google. The recommendation is to keep investing in traditional SEO while layering in AEO and GEO strategies on top of it.
What’s the difference between AEO and GEO?
AEO (Answer Engine Optimization) is about earning citations — getting your site linked as a source inside AI-generated answers like Google’s AI Overviews. GEO (Generative Engine Optimization) is about earning recommendations — becoming one of the brands ChatGPT, Claude, or Perplexity names when a user asks for options. AEO is earning the link back. GEO is being the recommendation itself.
Why did my organic traffic drop 20-30% and should I panic?
A meaningful portion of the value your site used to capture as visits is now being consumed without a visit — AI bots crawl your page, extract the information, and serve it inside ChatGPT or Perplexity without the user ever clicking through. A 30% drop in organic traffic does not mean a 30% drop in revenue. People are still buying; they’re discovering you through channels your 2019 dashboards can’t measure.
How do I track whether my brand is showing up in AI search?
There are three layers worth building: custom referral tracking in Google Analytics to catch human visits coming from AI tools; an AI visibility platform like Profound or Otterly that runs hundreds of prompts across major LLMs and monitors brand mentions; and brand sentiment monitoring, since how people talk about your brand online directly shapes what LLMs say about you.
Why shouldn’t I just search my own brand on ChatGPT to see how I’m doing?
Because ChatGPT personalizes results based on your account history. Your account already knows you love your brand, so the results are skewed in your favor. Use incognito mode, logged-out sessions, or a visibility tracker that depersonalizes the testing to get accurate data.
What is an AI agent and why does it matter for my marketing funnel?
An AI agent is an AI system that doesn’t just answer questions — it takes action on a user’s behalf. Instead of asking ChatGPT how to buy running shoes, a user tells an agent to actually buy them. The agent completes the transaction without the user visiting your site, which removes the website visit, the add-to-cart click, and the abandoned cart email opportunity from the funnel entirely.
Want help making sense of where your brand stands in AI search, and where to focus next?
That’s exactly the kind of audit our Collective is built to deliver.
Get in touch: https://apiarydigital.com/contact/
