How ChatGPT and other generative AI are changing search

Updated October 2026. We wrote this post in July 2023. The section “What’s changed since we wrote this” brings it up to date.

You’ve probably heard about ChatGPT and other large-language-model generative AI that has been all over the news recently. Depending on what business sector you are in, this may or may not be something you’re actively discussing at work. But one thing seems certain: with its ability to understand and generate human-like text, these AI tools will have huge impacts on nearly every business in the coming years. This is a technology with a very large potential to change the way we interact with the world.

One aspect of life where AI seems all but sure to change the existing model is with internet search. So, in this article we dive into the subject of AI and search, looking at the different ideas and predictions for how AI will change how we use the internet. By learning about these changes early, we can all begin to prepare for the coming landscape, putting ourselves in a better position to have continued success online.

What’s changed since we wrote this

In 2023, AI search was mostly a prediction. By October 2026, it is part of everyday searching:

• Google AI Overviews: the Search Generative Experience described below grew out of Google’s Search Labs into AI Overviews, AI-written summaries with links to sources shown above the regular results. Google began rolling them out to everyone in the US on May 14, 2024.
• Google AI Mode: a separate tab in Google Search for longer questions and follow-ups. After testing it in Search Labs, Google rolled it out across the US from May 20, 2025.
• ChatGPT search: OpenAI added web search with links to sources on October 31, 2024. It reached all logged-in users on December 16, 2024, and everyone, with no sign-up, on February 5, 2025.
• Perplexity: an AI answer engine that searches the web in real time and gives each answer numbered citations linking to its sources.

These answers take some of the clicks that used to go to websites. The Pew Research Center tracked the browsing of 900 US adults in March 2025. When Google showed an AI summary, people clicked a regular search result on 8% of visits, against 15% when there was no summary.

What hasn’t changed is what these tools rely on: web pages they can crawl, understand and trust. Google says its AI features have no extra requirements beyond the basics of good SEO. Our guide to optimizing your website for AI search turns that into a practical checklist.

The Continued Primacy of Search Engines

It’s helpful to remember where we are now and where we’ve been. Search engines and search engine optimization have been around in more or less the same form since the late 1990s. Granted, Google keeps making changes and improvements to its algorithm that marketing teams and developers must learn to account for. Over these years, Google and other search engines have introduced things like structured data, local packs, featured listings and more. But it’s also remarkable that in the last 20-some years, the way that we search has remained remarkably stable. We go to a search engine (nearly always Google since around 2004) and type a search query, we then scroll through the results. Recent AI tools present the possibility that all of this will be turned on its head.

Some of the ways that new generative AI tools stand to change the search environment.

Natural Language Understanding

Traditional search engines have relied on keyword-based queries, which often require users to craft their queries in a specific manner to obtain relevant results. New AI tools like ChatGPT, on the other hand, excel at natural language understanding. These tools can comprehend queries in a conversational manner, allowing users to ask questions in a more intuitive and human-like way and give answers that take context and tone into account.

Conversational Search Experience

The new AI tools bring a conversational element to search interactions. Instead of receiving a static list of search results, users can engage in dynamic conversations with the AI model. They can ask follow-up questions, seek clarifications, and receive instant responses. This conversational search experience enhances user engagement and enables a more interactive and personalized exploration of information.

Contextual Recommendations

The new AI tools’ ability to comprehend context goes beyond individual queries. They can retain information from previous interactions within a session, allowing for a more personalized and context-aware search experience. As users engage in ongoing conversations, these AI tools can provide tailored recommendations, suggestions, and insights based on the accumulated knowledge. This contextualization leads to more relevant and meaningful search results.

Handling Complex Queries

The new AI tools excel at handling complex and nuanced queries that might be challenging for traditional search engines. Users can pose multifaceted questions, seek explanations, or request comparisons, and these tools can process and generate comprehensive responses. Their deep understanding of language and ability to generate coherent and informative answers make them valuable tools for addressing intricate search queries.

Assisting with Decision-Making

Beyond providing search results, the new AI tools can assist users in decision-making processes. Whether it’s choosing between products, planning a trip, or seeking advice, these tools can offer personalized recommendations based on user preferences and specific criteria.

Enhancing Accessibility and Inclusivity

The new AI tools have the potential to make search more accessible and inclusive. Their conversational nature allows users with varying levels of technological literacy to interact with search engines more effectively. People who struggle with typing or navigating traditional search interfaces can simply express their queries in natural language, simplifying the search process and expanding access to information.

What the New Search Might Look Like:
Search Generative Experience (SGE)

This section is from 2023. SGE has since become AI Overviews and AI Mode, as described above.

Google’s new approach search has been dubbed the “search generative experience” and it differs from a “search query” in important ways. The major changes are related to what kind of information we expect to get from a search.

How We Search

Before, we would ask about “good seaside towns to visit in Greece” and we would get some snippets of text directly taken from a tourism site, plus a number of links to sites that list destinations and reasons for visiting them.

Now, we could ask for more specific information and get more nuanced responses. “What is a relatively inexpensive seaside town in Greece with renowned food that is serviced by train or bus from Athens?” With the new SGE, this would result in a response that the AI model generates itself, directly answering the question, with links to more information. Google would use bus and train maps, yelp reviews for restaurants, and other directories to make tailored recommendations for your trip, and would furnish links to hotels, train schedules and more. Instead of giving you the tools to answer the question for yourself, SGE tries to answer the question and then gives you tools to act.

Comparisons

Another example: recently, I searched for “new sneakers summer 2023” to see a list of the new sneakers released by the major brands, which I perused photos of as “Google Shopping” results. But with the new SGE my search would look different: “give me a comparison of new sneakers for long-distance running, that are good for people with low arches.” The results would be more tailored to my specific situation, and give me the precise information I’m looking for that may not already exist on the web but can be created by Google generative AI.

What You Can Do To Prepare for the Changes to Search: From Keywords to Content Depth

So what should you be doing to stay relevant in the new “SGE” internet? The key, many experts are saying, is a movement away from keywords towards “content depth”. Let me explain.

In the above example, notice how the search query goes from being largely keyword-based — seaside town, greece, etc. – to being context-based. The old search engines, for example, don’t understand that you are planning a trip, they might think you are just gathering information. The new search engines are more like digital assistants in that they can understand the intention behind the query, asking you questions to specify your particular needs, and then provide context-rich content.

This means that where we used to use keywords to gain search position and relevance with Google, we now need to go further and provide nuanced, in-depth content in order to be noticed by Google. Showing off what you care about, what you know about, and the different ways that it could be relevant is the new content marketing. Not only does AI gobble up that content, but also it remains important to have a human voice, to be a trustworthy source of information when AI-generated information is free and available. All this to say that where we used to create content, now we should be creating “deep content” – content that goes beyond the first questions and generic answers to get to the heart of nuanced, context-specific subjects. To put this another way: write content that can’t be written by AI.

What NOT To Do

Some clients have gotten in touch recently with questions like: should I just get ChatGPT to write my website, or my product descriptions, or the articles that I publish on my website? We’d still say no to publishing it as it comes out. Google’s guidance on AI-generated content (February 2023) says it rewards original, high-quality content however it is produced, and that using automation, including AI, mainly to manipulate rankings breaks its spam policies. In March 2024, Google added a scaled content abuse policy against mass-producing pages to rank, whether a person or a machine writes them. So what matters is whether a page is useful. Using generative AI to write a draft of an article or product description can be fine, as long as someone who knows the subject checks it, adds what only you know, and makes sure it sounds like you.

Some Final Thoughts

It’s clear that generative AI will change search and the way we use the internet in drastic ways. But websites with valuable, nuanced and unique information will always be what the internet runs on. The imperative to be relevant and bring depth to your subject has never been more important.

See our SEO and AI search work, or get in touch if you need help building a unique and valuable site, or have questions about ChatGPT and how it might affect your website.