How AI search engines differ
from Google

ChatGPT Search. ChatGPT activates web search for roughly a third of all queries. For the remaining two thirds, it generates an answer based on training data without consulting the web. When ChatGPT does search, it uses the Bing index and its own web crawling via OAI-SearchBot. Source selection focuses on narrative depth and expertise. ChatGPT prefers to cite sources that cover a topic extensively and with clear expertise. Short, superficial pages are rarely cited, even if they rank well in Google. For a solicitor in Birmingham, this means a detailed article explaining the conveyancing process step by step has a far better chance of being cited than a thin service page with just an address and phone number.

Google Gemini. Gemini has direct access to Google's own search index, the most extensive web crawl in the world. That gives Gemini an information advantage over other AI platforms. Source selection is based on Google's own ranking signals, supplemented with AI-specific criteria such as semantic completeness and answer quality. If you rank well in Google, you have a better starting position for Gemini, but it is not a guarantee of citation. Gemini is growing rapidly in the UK and is increasingly integrated into Google's core search experience.

Perplexity: the most transparent platform

Perplexity. Perplexity searches the web for virtually every query. It is the most transparent AI search platform: every answer contains numbered citations to sources. Perplexity has a strong preference for earned media (news articles, industry reviews, independent analyses) and a pronounced recency bias. Content published or updated in the past 30 days gets significantly more chance of citation. Perplexity is therefore the most comparable to a traditional search engine, but with AI-generated summaries. For UK estate agents or financial advisers publishing regular market updates, Perplexity is particularly relevant.

Google AI Overviews. AI Overviews are not a standalone platform but a layer on top of Google's search results. They appear for a growing proportion of search queries. Source selection is based on Google's existing index, with extra emphasis on semantic completeness and structured content. Google AI Overviews prefer to cite pages with clear sections of 200 to 400 words that answer a specific question. The click-through rate drops significantly when an AI Overview is shown, but sources that are cited in the Overview actually receive more clicks than they would have without it.

The most striking difference between these platforms is the overlap in sources. Only 11% of domains cited by ChatGPT are also cited by Perplexity. That means each platform selects largely different sources. A website that appears prominently in ChatGPT answers can be completely absent from Perplexity, and vice versa. This makes multi-platform monitoring essential. You cannot infer from one platform how you score on others.

A crucial difference from Google that many business owners do not know: AI search engines do not always search the web. Google searches its index for every query. ChatGPT activates web search for only about a third of queries. For the remaining two thirds, it generates an answer from training data: information the model learned during its training phase. That training data can be months or even a year old.

That has direct consequences for your visibility. If someone asks ChatGPT "What is a good strategy for AI visibility?", ChatGPT can answer from training data without visiting your website. Your recent blog post about AI visibility will not come into view unless ChatGPT decides to activate web search. The platform makes that choice based on the question: does this require current information or can it be answered with existing knowledge?

Perplexity: always live search

Perplexity sits at the other end of the spectrum. It searches the web for virtually every query. That makes Perplexity the most comparable to Google in terms of freshness. If you publish an article, Perplexity can cite it within hours. With ChatGPT, it can take weeks before your content is incorporated into training data, and even then web search needs to be activated to find your most recent content.

Google Gemini sits in between. It has access to Google's live index and can retrieve current information, but combines this with its own AI knowledge. Gemini is better than ChatGPT at integrating recent information, but less consistent than Perplexity in citing sources.

Google AI Overviews use Google's live index and are therefore always based on currently indexed content. That makes AI Overviews the most predictable: if your page is in Google's index and relevant to the query, you have a chance of being cited. The challenge is that AI Overviews select content differently from the traditional search results. You can be in position one in the organic results but not appear in the AI Overview, and vice versa.

Two strategies at once

The practical implication: you need two strategies. A long-term strategy to get into the training data of AI models (by consistently publishing high-quality content that gets widely cited), and a short-term strategy to appear in real-time search results (by publishing current, well-structured content that AI crawlers can retrieve). A Bristol-based marketing agency, for instance, would benefit from publishing authoritative guides about digital marketing trends (for training data inclusion) while also producing timely pieces on algorithm updates or platform changes (for real-time citation by Perplexity).

The UK market is shifting fast. Google still dominates with 93% search market share, but that number has been declining. Meanwhile, 54% of British adults now use AI tools, up from 31% just a year ago. ChatGPT alone received 1.8 billion UK visits in the first eight months of 2025. The UK is one of ChatGPT's top five global markets. These are not fringe users. These are your customers.

For a UK SME, this means a dual strategy. You need to maintain your Google SEO: that is still your largest traffic channel. But you also need to invest in AI visibility, because that is the growing channel. Those two strategies overlap partially (quality content, technical optimisation, authority building) but also have their own distinct requirements.

Example: a solicitor in Manchester

Consider a solicitor in Manchester. For Google, they have been optimising for years: local SEO, Google Business Profile, content marketing around legal topics. For AI visibility, they need extra steps. Their content needs to be structured so that AI platforms can extract direct answers. Their expertise must be explicitly visible: author information with qualifications, concrete case descriptions, practice areas with clear specialisation. They need to be listed on the Law Society Find a Solicitor directory, have Trustpilot reviews and maintain an active LinkedIn presence. And they need to monitor whether and how AI platforms mention their firm when someone asks for legal advice in Manchester.

An online shop has different priorities. Google shows product results via Google Shopping and organic results. AI platforms give product recommendations differently: they compare options, name pros and cons and give direct recommendations. An online retailer that wants to be cited by AI needs extensive product information: specifications, comparisons with alternatives, honest reviews and clear pricing. A thin product page with just a photo and price will not be cited. Retailers like John Lewis and Argos already structure their product pages with detailed specs, comparison features and buyer guides. Smaller e-commerce businesses need to follow this approach.

SaaS companies benefit most from the shift. Their target audience are early adopters who already use AI tools for research. According to recent data, 65% of B2B buyers use AI tools before contacting vendors, rising to 75% for those aged 25 to 40. A SaaS company that publishes extensive, current documentation and comparison content gets picked up faster by Perplexity and ChatGPT. Technical documentation that serves as reference material is particularly valuable: AI platforms cite this type of content frequently because it is factual and verifiable.

Expertise wins over optimisation

The common lesson across all sectors: Google rewards optimisation, AI rewards expertise. Those two are not mutually exclusive, but the emphasis shifts. A website that is purely optimised for Google rankings but has superficial content loses ground in the AI world. A website with deep expertise that is also technically well built scores in both worlds. Fewer than 10% of UK firms currently have a strategy for appearing in AI search results. That window of opportunity will not stay open much longer.