Search Engine Recap - October 2025

This month, we cover the latest in AI search across platforms, how links influence visibility in AI interfaces, and what Google’s Q3 numbers reveal. Here is an overview.

AI Overviews now overlap 54% with traditional results

New data shows that Google’s AI Overviews (AIO) now overlap with the top organic results in 54% of cases - up from around 32% at launch. That means 46% of AIO answers still come from pages that don’t rank organically for the same queries, pointing to a parallel retrieval path.

AIO is getting closer to traditional ranking patterns, but still shows independent behavior in how it selects content.

This could be driven by systems that prioritize semantic relevance over traditional link signals - meaning non-ranking pages can still surface.

The level of overlap varies by industry (eCommerce still lags), so impact is sector-specific.

For SEO, you need to win on both fronts: traditional rankings and AIO citations. Make answers scannable, use clean information architecture and schema, and keep facts fresh - semantic clarity is your key to visibility and authority.

New LLM study: Users want concise, actionable answers - and credible sources

A new study on real-world LLM behavior shows that users are leaning into chatbots for quick wins - short answers, step-by-step guides, and help with emails, social posts, or summaries.

Sessions are typically short and focused. Users prefer concrete examples and templates over long-winded explanations. Trust increases when responses cite recognizable sources, and drops when answers feel vague or generic.

For businesses, this means content should be:

• Scannable with clear headlines

• Supported by structured data

• Designed for reuse by LLMs (e.g. calculators, checklists, copy snippets)

Fresh, accurate content helps reduce AI hallucinations and increases citation likelihood. Success here isn’t just about clicks—it’s also about brand mentions and assisted conversions.

Microsoft revamps Bing Places + adds data-nosnippet support

Microsoft has relaunched Bing Places for Business with a new interface, recommendation engine, faster import from Google Business Profile, and automatic migration for existing users.

All listings remain free and manageable via bing.com/forbusiness, with dashboards, bulk edits, and real-time status updates making local SEO easier.

At the same time, Bing now supports the data-nosnippet HTML attribute, letting businesses exclude specific sections from snippets and Copilot AI responses -while still allowing the page to be indexed. A great way to protect paywalled content, UGC, or test pages.

Agencies and multi-location brands especially benefit from better import tools and bulk workflows. Developers should implement data-nosnippet selectively and monitor impact through Bing Webmaster Tools.

Google rolls out Lighthouse 13 with insights-first auditing

Google has launched Lighthouse 13, featuring a new insight-based auditing system that aligns with Chrome DevTools and removes low-value checks - without affecting performance scores.

Now live via npm and Chrome Canary, and rolling out to PageSpeed Insights and Chrome stable soon.

Audits are grouped into new “insights,” like:

• cls-culprits-insight

• document-latency-insight

• image-delivery-insight

Lighthouse 13 also includes metrics like lcp-discovery-insight and interaction-to-next-paint-insight (INP), improving focus on perceived performance.

Removed audits include font size, offscreen images, and font preloading. These aren’t direct ranking signals, but still matter for UX.

Google debuts Speech-to-Retrieval (S2R) for Voice Search

Google has unveiled a new Speech-to-Retrieval (S2R) system for voice search. Unlike traditional methods, it bypasses transcription, mapping speech directly to embeddings that are matched in semantic space.

A dual-encoder system processes both voice and text, enabling it to understand paraphrased queries without requiring keyword matches.

A secondary ranking layer then blends semantic matches with traditional SEO signals - so core SEO practices still apply.

Test results show S2R is faster, more accurate, and better across multiple languages. For brands, this means it’s critical to:

• Clearly address user intent

• Use strong site structure

• Demonstrate authority to increase semantic voice visibility.

New analysis: Links do matter for AI visibility - surprise: nofollow too

An analysis of 1,000 domains and 35,000+ data points found that backlinks do play a role in AI visibility - but quality beats quantity.

Authority (measured by tools like Semrush’s Authority Score) correlates with AI Share of Voice, especially among top-performing domains.

The twist? Nofollow links correlate nearly as well as follow links - suggesting LLMs treat them equally when determining credibility. Image links also drive strong results, particularly for mid-to-high authority pages.

TL;DR:

• Focus on link diversity

• Don’t ignore nofollow or image-based links

• Prioritize high-authority, relevant sources

AI systems reward a rich, credible link profile, not just volume.

Google’s Q3: AI tools are complementing - not replacing - Search

Google’s Q3 earnings reveal that AI Mode and AI Overviews are additive, not a replacement for Search. Both total queries and commercial searches increased year-over-year, with growth outpacing Q2.

AI Mode queries have doubled, now reaching 75 million daily users across 40 languages. Alphabet recorded its first $100B+ quarter, driven by Search and YouTube Ads. Shorts now earn more revenue per watch hour than in-stream ads.

Despite a $3.5B EU fine, Alphabet maintained a 33.9% margin, leaving plenty of room for further AI investment.

Google also teased Gemini 3, new Chrome integrations, and a larger 2025 CapEx budget - clearly signaling more AI-first innovation is coming.

Takeaways & Recommendations

October brought updates from both Google and Bing - alongside fresh insights into LLM behavior.

What’s clear is that Search - and therefore SEO - is becoming multi-layered. That means more opportunities for visibility and growth, but also more complexity in how we optimize.

Here’s what to focus on:

• AI Overviews are aligning more with organic rankings - but you still need to optimize for both.

• Structured, actionable content is a must for LLM visibility. Think clarity, trust, and utility.

• Lighthouse 13 is your signal to double down on performance, especially around INP and Core Web Vitals.

• Bing Places has had a major upgrade—well worth optimizing for local visibility (and potentially ChatGPT integrations).

• And finally, link diversity and authority matter more than ever, even for AI visibility.

Not sure how to best implement these changes across your business?

Reach out to Bonzer. As a specialized SEO agency, we help ambitious companies win top visibility in both classic and AI-powered search - every single day.