Google has shipped several updates this week that directly affect how we track organic performance for our clients. Two Search Console changes alter the way we measure traffic from AI Overviews and image-based searches. Meanwhile, Google’s broader product ecosystem — from Gemini’s new text-to-speech models to Google Photos and Health app updates — signals a continued push toward multimodal, AI-first experiences. Here is what matters and what we are already implementing.

Key Takeaways

• AI Overview links now split between website clicks and AI Mode entries, and Search Console tracks each differently.
• A new multimodal filter in Search Console isolates traffic from image-based searches — but strips query data.
• Gemini 3.8 Flash TTS models introduce highly expressive text-to-speech, pushing content accessibility expectations forward.
• Google Photos updates reinforce visual search and on-device AI as core user behaviours brands must optimise for.
• Google Health app features on Pixel Watch show the expanding surface area of Google’s ecosystem beyond traditional search.

AI Overview Links Now Route Two Ways — Track Accordingly

Google’s AI Overviews are displaying more links than before, but not every click lands on a website. As Search Engine Journal reports, some AI Overview links now route users into AI Mode rather than to external sites. Search Console measures these interactions differently, meaning the click data our clients see may not reflect actual site visits.

We are auditing Search Console reports across all client accounts to separate genuine referral clicks from AI Mode redirections. If you rely on AI Overview visibility as a traffic metric, you need to recalibrate expectations immediately. Raw impression counts in this context are misleading without understanding where the user actually ends up.

Search Console’s Multimodal Filter Reveals Image-Search Traffic

Google has quietly added a multimodal filter to Search Console, and it is a significant data point for any site with visual content. According to Search Engine Journal’s coverage of the new multimodal filter, the feature shows which pages receive traffic from image-based searches. The catch: query data is not included.

For our e-commerce and portfolio clients, this changes how we evaluate image SEO performance. We are cross-referencing multimodal traffic data with existing image sitemap submissions and structured data markup. Pages receiving image-search traffic without proper alt text, schema, or optimised file names represent low-hanging wins we can act on today.

Gemini 3.8 TTS Models Raise the Bar for Audio Content

Google has launched Gemini 3.8 Flash-Lite TTS and Gemini 3.8 Flash TTS, described as their most expressive text-to-speech audio models to date. The quality gap between synthetic and human narration is narrowing fast.

For our clients producing long-form content, this opens a practical route to audio versions of blog posts, guides, and product descriptions at minimal cost. We are testing these models for automated podcast-style summaries of client content. Accessibility compliance also benefits — high-quality TTS removes friction for users who prefer or require audio consumption.

Google Photos and Visual AI Habits Shape Future Search Behaviour

Google has rolled out five new Google Photos features designed around visual memory management. These updates reinforce a pattern: users increasingly interact with Google through images, not just text queries.

This matters for SEO because visual search habits trained in consumer apps migrate to Search. We are advising clients to treat image optimisation with the same rigour as on-page copy. Proper EXIF data, descriptive filenames, and contextual surrounding text all feed into how Google indexes and surfaces visual content.

Google’s Expanding Ecosystem Means More Surfaces to Monitor

The launch of new health and safety tools in the Google Health app for Pixel Watch is not directly an SEO story, but it illustrates a critical trend. Google’s ecosystem now spans wearables, AI assistants, visual tools, and traditional search. Brands operating in health, wellness, or lifestyle sectors need to consider how their content appears across these surfaces — not just in the ten blue links.

The common thread across all five updates is clear: Google is fragmenting the ways users discover and consume information. Our team is responding by diversifying how we measure visibility, optimise assets, and report results. Sticking to traditional click-through metrics alone will leave gaps. We recommend every site owner review their Search Console multimodal data, audit AI Overview link behaviour, and begin testing AI-generated audio content before the end of Q3 2025.

Frequently Asked Questions

What is the new multimodal filter in Google Search Console?

It is a filter that isolates pages receiving traffic specifically from image-based searches. It shows which pages benefit from visual search but does not currently provide the specific queries users typed.

How do AI Overview links affect website traffic in 2025?

Not all AI Overview links send users to external websites — some now redirect into Google’s AI Mode instead. This means Search Console click data from AI Overviews may overstate actual site visits unless you account for this split.

Why does image SEO matter more now for Google rankings?

Google’s new multimodal filter confirms that image-based searches drive measurable traffic to websites. Optimising alt text, file names, and structured data for images directly improves visibility in this growing search channel.

What is Gemini 3.8 text-to-speech and how can websites use it?

Gemini 3.8 TTS is Google’s latest AI model for generating natural-sounding audio from written text. Websites can use it to create audio versions of articles, product descriptions, and guides — improving accessibility and user engagement at scale.