A DTC skincare founder emailed us last month with a spreadsheet. Organic traffic down 34% year-over-year. Panic mode. She was ready to fire her SEO agency and put everything into "AEO" because a LinkedIn thread told her Google was dead.
We pulled her data. Her informational blog traffic had cratered, sure. That's the "how to build a skincare routine" stuff. But her product pages? Up 12%. Her category pages? Flat. Her branded search? Up 40%, because ChatGPT was starting to recommend her by name and people were Googling to buy.
Google isn't dying. It's specializing. And AI search isn't a replacement. It's a different job.The founders who lose the next two years will be the ones who treat this like a migration. The ones who win will treat it like a portfolio.
The Query Split Nobody's Drawing Cleanly
Here's the shift, in the plainest terms:
Google is becoming a transactional engine. People still Google when they want to do something. Buy a supplement. Book a plumber. Compare pricing. Click a link. AI is becoming a research engine. People ask ChatGPT, Claude, and Perplexity when they want to understand something. Which supplement is best for postpartum hair loss? How do I choose between two CRMs? What are the trade-offs of switching to headless commerce?The data backs this. According to Similarweb's Q3 2025 report, Google's informational query volume dropped 21% year-over-year, while transactional query volume (queries with commercial or navigational intent) grew 6%. Meanwhile, ChatGPT's referral traffic to publishers grew 12x in the same period, but the sessions were short and research-shaped — high time-on-page for editorial content, near-zero conversion on product pages when they even bothered to click through.
Translation: the informational middle of the funnel is migrating to AI. The bottom stays with Google.
That's not a headline. That's an allocation instruction.
What Actually Happens on Each Channel
Let's get concrete. A shopper is considering buying a new mattress. Here's the split in 2026:
She asks ChatGPT: "What's the best mattress for a side sleeper with lower back pain?" ChatGPT gives her three brands with reasoning: Saatva, Helix, Nectar. She reads. She thinks. She's not clicking anything.
She goes to Google: "Saatva Classic queen price". She clicks. She buys.
The AI recommendation happened first. The transaction happened on Google. If you're Saatva and you're only measuring one of those two moments, you're missing half the picture.
Now flip it. A B2B buyer is evaluating email platforms.
She asks Claude: "Klaviyo vs Attentive for a $2M ARR Shopify brand?" Claude gives her a breakdown. She reads for six minutes. She never clicks.
She Googles: "Klaviyo pricing". She books a demo.
Same pattern. AI does the shortlist. Google does the transaction.
Being on the shortlist is the new ranking. Getting the click is still the old ranking.The Migration by Query Type
Here's how it's shaking out across the four major query intents, based on aggregated data we've pulled across ~800 Foxish customer accounts in 2025:
| Query Type | Example | Where It Lives in 2026 | Volume Trend |
|---|---|---|---|
| Informational (broad) | "how does creatine work" | AI (ChatGPT, Perplexity) | Google -38% YoY |
| Comparative | "Notion vs Coda" | AI-first, Google secondary | Google -24% YoY |
| Navigational | "asana login" | Flat | |
| Transactional | "buy Bombas socks" | +9% YoY | |
| Local | "dentist near me" | Google (AI weak here) | +4% YoY |
| Reputation-checking | "is X supplement legit" | AI-first | Google -19% YoY |
| Long-tail specific | "wireless earbuds under $80 with ANC" | AI-first | Google -31% YoY |
First, anything that requires synthesis is going to AI. Comparisons, "best of" lists, nuanced recommendations, reputation checks. These are the queries where Google used to send you to 10 blog posts and you had to piece it together. AI does the piecing for you now.
Second, anything that requires action stays with Google. Buying, logging in, finding a location, booking. AI doesn't handle commerce well yet — no reliable cart, no reliable local database, no persistent identity across platforms.
The middle got hollowed out. The edges are stronger than ever.
Where SEO Still Wins (And Isn't Going Anywhere)
Product pages are still a Google game. When someone searches "Bombas ankle socks black," they want a PDP, a price, and an add-to-cart button. ChatGPT doesn't do that. Perplexity doesn't do that. AI Overviews sometimes surface a product, but the click still lands on your Shopify page, and ranking in the top 3 organic positions on Google is still what determines whether they buy from you or a reseller.If you run ecom, this is where 60-70% of your SEO investment should go: product schema, PDPs, category pages, review markup, site speed, and the boring blocking-and-tackling that's been the job since 2015.
Local intent still lives on Google. "Coffee shop open now near Bushwick" is a Google Maps query. It will remain a Google Maps query for the foreseeable future. AI doesn't have the local data infrastructure, and even Google's AI Overviews defer to Maps for these queries. If you're a service business, your Google Business Profile is still 50% of your traffic strategy. Navigational and branded search stays on Google. When someone knows your brand name, they Google it. They don't ask ChatGPT "take me to Warby Parker." Branded search is where AI visibility actually converts — AI creates the awareness, Google captures the intent. Your job is to make sure your branded SERP is clean, complete, and conversion-optimized. Sitelinks, knowledge panel, review stars. High-commercial-intent long-tail still converts on Google. "Best CRM for real estate agents under $50/month" might get answered by ChatGPT for research, but the buyer who's ready to trial ends up on Google typing your product name. That's where they land on a comparison page or a pricing page. Your G2, your Capterra, your review-heavy landing pages — still Google real estate.Where AEO Wins (And Where SEO Can't Compete Anymore)
Discovery for anything new or unfamiliar. If your brand is under three years old, if you're in an emerging category, if you don't have the domain authority to rank against Wirecutter and Forbes on Google, AI is your fastest path to being named. We've watched brands go from zero mentions to being cited in 40% of relevant ChatGPT queries in 90 days by publishing structured, Q&A-shaped content on their own domain and getting cited in 15-20 mid-tier publications.That's not possible on Google. On Google, that trajectory takes 18 months and a $200K content budget.
Comparison and "best of" content. This is where AI has completely displaced Google. In 2023, ranking for "best project management software" was a viable strategy. In 2026, that ranking gets you 40% of the traffic it used to, because ChatGPT is answering the same query without a click. The play now isn't to rank the listicle. It's to be in the listicle that ChatGPT builds on the fly. Trust and reputation queries. "Is this company legit," "does this actually work," "any complaints about X" — these are AI queries now. And the answer AI gives is based on Reddit threads, forum posts, review aggregators, and (increasingly) your own on-site FAQ and customer story content. If you don't own the answer to "is [your brand] legit," you're letting a random 2019 Reddit thread define you. Complex B2B evaluation. SaaS buyers now spend 6-8 minutes with an AI before they visit a single vendor site. If your product isn't named in that conversation, you're not on the shortlist. It doesn't matter how good your Google Ads bid is if you weren't recommended in the research phase. The old funnel was: search, click, evaluate, buy. The new funnel is: ask, get recommended, search by name, buy.Notice the shape. AI at the top. Google at the bottom. Both matter. Neither is optional.
How the Two Channels Compound
Here's the part most operators miss: AEO and SEO don't compete. They compound.
When ChatGPT recommends you, three things happen. First, the person may click through directly (small percentage, but real). Second, and much bigger, they Google your brand name. Third, they remember the name for later.
Branded search is the compounding mechanism. It's the receipt of AI visibility. Every time your AI mentions go up, your branded Google search volume goes up two to six weeks later. We see this pattern in almost every account we track. A skincare brand we work with went from 340 branded searches per month in March to 2,100 in September — and the trigger was a 4x increase in ChatGPT mentions between February and June.
That branded search is high-intent, low-CAC, and converts at 4-8x the rate of cold non-branded traffic. It's the best traffic you can buy. Except you're not buying it. You earned it by being named upstream.
Which is why treating AEO and SEO as separate line items in your budget misses the interaction effect. The right frame is: AI creates the awareness, Google captures the demand. You can't do just one and expect the funnel to work.
The founders we see fail are the ones who go all-in on one side. All AEO, no SEO: they get mentioned, but they have no product pages ranking for their branded terms, so a reseller or Amazon listing captures the transaction. All SEO, no AEO: they rank fine, but they're invisible in the research phase, so they never get shortlisted for anything they don't already own the branded search for.
Allocating Budget by Channel-Fit
Here's the practical part. If you're an ecom brand doing $2-20M in revenue, this is roughly how the split should look in 2026:
~50% of content investment goes to SEO fundamentals. Product pages, category pages, review markup, site speed, technical SEO, and Google Business Profile if you have any local component. This is your transaction infrastructure. It's boring. It works. ~30% goes to AEO-specific content. Q&A-shaped pages, comparison content on your own domain, structured data (FAQPage, Product, HowTo schema — LLMs use it), and Reddit/forum/community presence. Your goal here is to be the source AI cites, and to be mentioned in the third-party content AI reads. ~20% goes to reputation and PR. Getting cited in publications that AI trusts. Wirecutter, industry pubs, respected newsletters. AI weighs source credibility, and this is how you buy your way onto the credible list. A single feature in the right publication can move your ChatGPT mention rate more than 30 blog posts on your own site.For B2B SaaS, the split skews harder toward AEO — maybe 30% SEO, 40% AEO, 30% reputation — because buyers spend more time in the research phase and your product pages matter less than your presence in comparison conversations.
For local services, flip it. 70% SEO (Google Business Profile, local citations, review velocity), 20% AEO, 10% reputation. AI still doesn't handle local well.
The Measurement Problem Nobody's Solved
This is where most operators get stuck. You can measure SEO. You can measure AEO. But almost nobody is measuring the interaction between the two.
Google Search Console tells you rankings and clicks. Fine. AI visibility tools (including ours) tell you mention rates and share of voice. Fine. But the question that actually matters — is my AI visibility driving my branded search, and is my branded search converting? — sits in a gap that most analytics setups don't cover.
The proxy metric to watch is branded search volume plotted against AI mention rate, on a 4-8 week lag. If your mentions are climbing and your branded searches aren't following, something's broken. Maybe the AI is naming you but not driving memory. Maybe your brand name is hard to remember or spell. Maybe you're being cited alongside 8 competitors and you're getting drowned out.
If your mentions are climbing and branded searches are climbing but conversions aren't following, your Google-side infrastructure is broken. Your PDPs, your branded SERP, your checkout. That's an SEO problem, not an AEO one, but the AEO investment is exposing it.
The channels are separate on paper. In practice, they're the same funnel measured at two different points.
The Real Takeaway
Stop asking "should I do SEO or AEO." Ask "which query types matter for my business, and where do those queries live now."
If you sell things people buy on Google, you need SEO. That hasn't changed and won't change in 2026.
If you sell things people research before buying, you need AEO. Because the research isn't happening on Google anymore.
Almost every business needs both. The mix is what varies. And the mix should follow the query-type distribution of your actual customers, not the LinkedIn narrative of the month.
The founder we opened with, the one with the crashing organic traffic? We didn't tell her to fire her SEO agency. We told her to shift 30% of the content budget to AEO-shaped work and keep the rest on PDPs and technical. Six months later, informational traffic was still down, but branded search was up 60%, ChatGPT was mentioning her in 22% of relevant queries (up from 4%), and revenue attributable to organic + branded was up 18% year-over-year.
The channels compounded. That was the whole trick.
Foxish measures both sides of this funnel out of the box: your AI mention rate across ChatGPT, Claude, Gemini, and Perplexity, plotted against your branded search volume and Google visibility. So you can actually see the compound effect, not just guess at it. First scan is free, no credit card. Start a free scan →