June 2025 Core Update

Google clarified its AI playbook; here’s what matters for search teams

Liz Reid, the new head of Search, told Semafor that we’ve entered the “AI search era.” TL;DR organize the web, faster than before, and use Gemini reasoning on-the-fly.

More AI, more questions

• According to Google, AI Overviews now nudge people to fire 10% more multi-step queries. Those strings run two-to-three times longer than the average blue-link search, which gives Google richer intent in one shot; and squeezes mid-funnel clicks. We reported back in April that we were starting to see more and more searches per visit in Google Search via Similarweb clickstream data. More ad revenue for Google! Also keep an eye on Google earnings calls because they've been using "query growth" as a KPI recently (maybe longer).
• Early industry data shows CTR for top listings drops up to A LOT when an AI Overview appears.
• The measurable, browser-based web is already shrinking ~3 % a year, so every lost click hurts more.
• Meanwhile ChatGPT now sends users to external sites 2.3× more often than Google Search, but we're not sure what's an agent/bot and what's a human!

Longer queries + lower CTR = a traffic crunch nobody can ignore.

Google still sweats spam. Reid says they keep it “under one percent” and expect new black hat tactics around AI answers (cough, cough.... LinkedIn Pulse).

About E-E-A-T

Hours after the interview, Search Liaison Danny Sullivan repeated the mantra: E-E-A-T isn’t a ranking factor. We still use the framework at Momentic because it forces clear sourcing, proofs of experience, and content people want to cite; habits that drive the real signals (links, engagement, brand queries) Google does measure. Keep the compass, ditch the myth of a secret score.

Skeptical? You should be. But a massive rise in AI-overview queries and 3× longer user experience in Google Search both scream: broader, richer language is flowing into the index fast. Pages that target only head terms and "keywords" with measurable search volume will fade. (Opinion)

What to watch next

• Roll-out volatility. June’s core update is live and will rattle rankings for roughly three weeks.
• Search Live. Reid teased a voice-first “interactive Gemini” mode that floats web links beneath spoken answers. If it sticks, organic slots may reopen.
• Ad experiments. Google will test ads inside AI Mode. How heavy the ad load gets will hint at revenue confidence.
• More impressions, flat or falling clicks in Search Console.

The playbook

Double-down on influence and brand.

1. Surface lived expertise. Show credentials, primary sources, and firsthand examples. AI Overviews grab what looks most real.
2. Write for human and machine. Front-load answers. Use Semantic HTML, e.g. H-tags, bullets, short sentences, and schema. Give Gemini/other LLMs a pre-built summary.
3. Render server-side. If view-source can’t see it, neither can an LLM... except Gemini.
4. Log referrers at the edge. Catch “direct” traffic coming from AI assistants and reclaim the credit.
5. Track citations, not just sessions. Clickstream for macro shifts. GA4 / GSC for on-site truth. CRM or revenue for bottom-line proof. Influence surfaces and figure out a way to measure that..

Push refreshed, multi-step problem-solvers into the communities your audience already trusts; real-world engagement lines up the signals Google will crawl next.

Google spokespeople can deny an E-E-A-T knob all day. The behaviours that prove experience and authority still correlate with the signals that rank. Keep the framework.

Adapt now, or watch your competitors become the source for AI/LLM/Chatbot citations.

Stay calm. Find your lane. And, find your value.

• Sudden drops in organic clicks with stable rankings
• Appearance of AI Overviews above your search results
• Decrease in branded click-through rates
• Pages previously driving informational queries may now be summarized and may or may not reference your website
• Increase in Google Search Console impressions, but not clicks
• Gain/loss of visibility in passage-level citation within AI Mode (manual testing for now)

• Refocus content strategy around semantic completeness at the passage level
• Ensure content includes verifiable claims, clear value props, and entity-rich structure
• Use semantic HTML (H1s, bullets, FAQs) to increase snippet eligibility
• Target intent clusters (not just keywords) using query fan-out simulation
• Start to figure out how you're going to monitor citation frequency in AIOs, AI Mode, and other Chatbots/LLMs. There's plenty of tools out there, but make sure you understand how they work and what their measurement weaknesses are.

‍