How AI Overviews and the Rise of Conversational Search Are Redefining SEO
Google’s May 2024 announcement of AI Mode in Search and the expansion of AI Overviews marked a pivotal shift in how search works — and how SEO professionals need to think. If you’re a marketer or publisher relying on organic traffic, this isn’t just a technical update. It’s a fundamental change in how people search, how they interact with content, and whether they ever even visit your site.
Why This Matters Now
You might already be seeing it: organic traffic is plateauing or dipping, direct traffic is rising, and user journeys are harder to trace. Don’t blame bad tracking — it’s AI tools altering how content is discovered, processed, and used. In Google’s new AI Mode, retrieval-augmented generation (RAG) powers Overviews that surface direct answers and passages from indexed sites — all without a user needing to click through.
The age of zero-click search is here. And marketers need to adapt.
Are People Still Searching?
Yes — but not always on Google. Users are turning to:
- YouTube and TikTok for how-tos
- Instagram, WhatsApp and Reddit for recommendations and real talk
- ChatGPT for planning and decision-making
- Amazon, Pinterest, and niche apps for product discovery
Google is still dominant, but the shape of search has changed. Many queries now trigger AI Overviews. Mike King, Founder and CEO of iPullRank recently reported that around 20% of Google queries include AI-generated summaries — and this is just the beginning.
What Is an AI Overview?
AI Overviews are generated by Google’s large language models using RAG (retrieval-augmented generation). The model:
- Chunks your page content into semantically meaningful bits
- Indexes those bits for relevance
- Retrieves and synthesizes them into natural-sounding answers to user queries
The links still appear — but let’s be real: users don’t always click them. This is what SEO folks are calling the Agentic Internet: a future where humans interact with AI agents who retrieve and synthesize info for them.
Why Traffic Attribution Is Breaking
Most AI tools — including Google’s AI Overviews and ChatGPT — do not pass referrer data, meaning when users do click through, they show up as “Direct” traffic in your analytics. Worse, some AI bots don’t respect robots.txt and operate in ways traditional analytics tools weren’t built to track.
Until recently, AI Mode traffic was labeled with rel=”noreferrer noopener”, making it indistinguishable from true direct visits. As of May 28, 2025, Google announced a fix: this traffic is now tagged as Google / organic — but not all bots or tools have caught up.
How to Track AI Traffic
Two excellent guides to get started:
- How to Track Traffic from AI Overviews (AIO) — Dana DiTomaso’s setup with GA4
- Tracking AI Bots in GA via GTM + Looker Studio — Jori Ford’s SEO Week talk
These approaches involve creating custom triggers in Google Tag Manager that fire when specific user-agent strings or behavioral patterns are detected. Then, that data can be visualized in Looker Studio dashboards.
AI’s New Priorities: From UX to AX
Humans look for compelling headlines and good navigation. AI agents, on the other hand, look for:
- Semantic HTML
- Schema.org structured data
- Clean, structured URLs
- Authoritativeness and domain trust
Marketers now need to optimize for the Agent Experience (AX) as much as the User Experience (UX). This means ensuring your content is machine-readable, semantically structured, and easy to index.
So, What Should Marketers Do?
- Expect lower traffic from organic search.
Users will get answers before reaching your site — if they click at all. - Shift your focus from activity to outcomes.
You may not track every visit, but sales, leads, and calls still matter. Track what counts. - Embrace structured content.
Use schema, semantic markup, and well-organized data to feed both humans and machines. - Track what you can.
Set up GA4 and Looker dashboards to surface AI-related traffic patterns. - Stay critical.
Understand that AI answers aren’t always accurate — and users may stop verifying sources. - Keep learning.
Amazon’s RAG Explainer or intros to vector embeddings can deepen your understanding of how AI reads your site.
Resources Worth Exploring
- Wil Reynolds on Paid vs. Organic in AI Search
- Lily Ray’s SEO in the Age of Agents
- Jori Ford’s SEO Week Presentation (YouTube)
- Anil Dash: AI-First is the New Return to Office
- MIT Press: Can AI Write Authentic Poetry?
- Matthew Berman on We Finally Figured Out How AI Actually Works (8:07 How AI Plans a poem is particularly interesting)
Final Thought: Don’t Panic — Prepare
We’ve lost visibility before. Remember the rise of “not provided” keywords in Google Analytics where 99% of the keyword data disappeared? This is just the next evolution. Marketers who adapt — by prioritizing outcomes, structuring content, and understanding how AI agents behave — will be ready for the road ahead.
