AI Mode: Its arrival in the UK and what it means for search

Google’s AI Mode has gone from the Lab to the lands this year. Back in March it was just a tab in Google Labs, but has since been launched in the US and India (June) and from 28th July, is being rolled out across the UK.

AI Mode Timeline

7th March 2025: AI Mode tab released in Google Labs
20th May 2025: Announced AI Mode will be launching in the US outside of Labs
13th June 2025: Launched in the US
24th June 2025: Launched in India
28th July 2025: Launched in the UK

For the uninitiated, AI mode is Google’s response to OpenAI’s chatGPT having introduced conversational search to the arena. It’s essentially a product of forced innovation that Google’s CEO Sundar Pichai views as “a total reimagining of search”. The staged roll-out is helping Google to establish what works and what doesn’t before incorporating features into Google’s central experience.

Key Insights

As of mid-July 2025, AI Mode had reached about 100 million monthly active users in the US and India. With the UK rollout, and other countries on the way, this growth will continue to gain momentum.
Users engaging with AI Mode submit queries that are 2–3 times longer than traditional search queries. This indicates increased appetite for complex and conversational searches: leading to longer tail, prompt-based search queries.
The average AI Mode user session is reported at 4 minutes and 37 seconds, which is significantly longer than with classic search, reflecting deeper engagement.
AI Mode results are currently extremely inconsistent, the same query run multiple times a day can yield different results from differing sources.
On average, AI Mode search results present 12.6 links with most 90.8% appearing in separate blocks and 8.9% as inline text links.
The uptake in AI users clicking the AI Mode tab in the US sits at just 1% according to early June data.

Impact on search

StrategiQ’s SEO and Content team have been keeping a close eye on AI Mode since its Google Labs release and subsequent rollout in the US. Their views on why publishers and websites need to consider its impact carefully include:

This is more than a UI change in Google search. AI Mode will continue to be developed and will transform the traditional search landscape.
Google’s advantage is the personal context, it knows you, it’s connected to your life already, your inbox, your calendar… AI shopping is coming.
Websites are now sources. This needs to be considered when you are designing content, having a new website built, or appraising your current one.
AI Mode’s query fan-out expands a user query into multiple sub-queries to capture possible variations in intent. This presents an opportunity, so ensure that semantically relevant content is accessible across similar pages.
Building your brand footprint across all digital assets is hugely important. What others say about your brand across forums, social, video and third party websites continues to feed these answer engines.
Traditional search optimisation standards remain true. Google has always been about what is best for users, so this continues to sit at the forefront of a great search strategy.
Zero click searches are likely to increase, so brands and websites need to work harder to entice the clicks. One route is to consider what unique proposition or exclusive offer the user might be interested in.

AI Mode represents the gateway to personalised search and with Google’s backing, may become the future state of search. According to Gary Illyes from Google, the “web is thriving” with AI Mode, AI Overview and AI Search delivering higher quality traffic.

However, one recent study from Ahrefs, suggested the opposite. They looked at the traffic of ~82K websites between May and June 2025, to explore user behavior changes depending on where visitors come from. The outcome, AI visitors visit fewer pages and bounce more often than traditional search visitors.

The truth probably lies somewhere in between these two views. As Google continues to evolve, their core focus on users has always ensured that they benefit from search improvements, (albeit with slight imperfections on the way).

Search UI

The Google homepage will now show the AI Mode option as a clear CTA, next to voice and image search options.

AI mode isn’t the default results page yet, but is shown as the first tab in results pages. Maybe this is an indication that it’s primed to become the future default?

AI Mode is signposted when AI Overviews appear in SERPs – users are able to research more with AI mode via CTA’s.

Whilst there is a subtlety to the presence of AI Mode in SERPs, we believe that it won’t be long before it becomes more prominent.

AI Mode Usage Data

Google has recently confirmed that AI Mode clicks, impressions, and average position metrics are now incorporated into the standard Search Console Performance report under the “Web” search type, meaning AI Mode is treated like traditional search traffic rather than being separated out.

A click is counted when a user selects an external link presented within AI Mode.
Impressions are recorded if a page appears in an AI Mode display.
Positions follow normal search ranking rules, with each AI Mode component (such as link cards, carousels or image blocks).
Follow-up queries within the same AI Mode session are logged as new search queries, which resets metrics, meaning clicks, impressions and rankings for those responses are measured separately.

While this ensures all AI Mode traffic is visible in Search Console, it cannot currently be isolated or segmented from classic web search data, which complicates attribution and analysis for SEO strategies.

Within your search console accounts, you are likely to see an increase in impressions, similar to that experienced with AI Overviews.

We hope that Google will eventually separate out AI Mode data to help distinguish between traffic from classic search results and AI-generated responses.

Query Fan-Out

Powered by Gemini 2.5, Google’s most intelligent model, AI Mode ranks results in a different way to traditional Search. It uses a query fan-out technique, breaking down a query into subtopics, before searching on each, and then compiling an AI generated answer.

Query Decomposition – AI Mode analyses the original query and identifies its core components and underlying intents, breaking it down into several distinct sub-questions.
Parallel Information Retrieval – Once the sub-queries are generated, the AI system “fans-out” and simultaneously executes these queries across a vast array of data sources.
Answer Generation – After retrieving information from all these parallel searches, the AI system then synthesises the findings.

The query fan-out technique represents a significant evolution in how information is discovered and consumed online.

Multimodal & Deep Research

Google’s AI Mode introduces a new level of search sophistication by combining advanced reasoning with multimodal capabilities, allowing users to input text, voice, or images to receive context-rich results.

The Deep Search feature significantly enhances the traditional “fan-out” method by issuing hundreds of parallel searches to generate comprehensive, expert-level summaries in a report style, complete with cited sources.

Users can then refine their results by asking follow-up questions, enabling a more tailored and interactive search experience, mirroring the depth and flexibility of a real conversation.

Agentic Capabilities

At Google I/O earlier this year, they also began to preview agentic features that can take action on a user’s behalf within AI Mode.

These upcoming capabilities, introduced under Project Mariner, will allow users to delegate tasks such as booking tickets or calling local businesses directly through Search. While initially limited to the US and select user tiers, these agentic features mark a major shift from passive information retrieval to proactive task execution. Google’s roadmap suggests a future where Search doesn’t just find information, but completes multi-step tasks for you, edging closer to the ‘virtual personal assistant’ anticipated.

Impact on Publishers

Whilst AI Overviews have steered clear of responding to hard news queries, AI Mode takes a different approach that could impact traffic driven from search to publishers:

The same result in AI Mode:

This development is seen as a concern in the publisher space. We have seen the likes of Cloudflare champion publishers by introducing new ways to be rewarded for their content.

Matthew Prince of Cloudflare commented “If the Internet is going to survive the age of AI, we need to give publishers the control they deserve and build a new economic model that works for everyone.”

Cloudflare argues that AI breaks the unwritten agreement between publishers and crawlers. Publishers using Cloudflare are able to block AI bots from accessing content without permission. Eventually they’ll be able to ask for payment in return with their “Pay Per Crawl” features.

This represents a huge discussion point, as many LLMs have used publisher content to train their models, with very little given in return. Many believe that society is normalising the massive copywriting breaches that AI companies are benefiting from.

Spurred by the introduction of AI overview, the CMA are introducing a new regulation in October that gives publishers more control over the way their work appears in Google. With the release of AI mode, the CMA may need to apply additional pressure on Google.

However, it should be noted that Google is testing tools in the US like their “preferred sources” experiment. This allows users to personalise the news they see surfaced in SERP features like top stories, by selecting their preferred publishers.

Ecommerce Impact

AI Mode in the UK will transform how ecommerce websites are discovered in search, by pulling product details, comparisons, reviews, and social proof into an informative summary at the top of the results page. This makes it more important than ever for ecommerce brands to:

Keep product information accurate and up to date
Use effective schema markup
Ensure that visible trust signals like authentic customer reviews and recent third-party mentions are readily available for Google to feature.

This new AI-powered experience is also fundamentally changing user behaviour. Fewer organic clicks to ecommerce sites are likely, as more questions are answered directly within Google.

At the same time, the familiar repetition of similar website links and paid ads is being reduced. This creates a cleaner, more helpful search environment centered around the user’s journey and intent. It also often keeps users on Google longer, building confidence towards a purchasing decision.

For ecommerce marketers, this means that standing out now requires not just traditional SEO, but a holistic strategy blending authoritative, well-structured content, genuine trust signals, and a deep understanding of how AI interprets and aggregates web information.

Example search: Mens waterproof jackets