The New AI Platforms That May Replace Search Engines

Sam Altman, the CEO of OpenAI, the company that created ChatGPT, almost casually mentioned that he had stopped using Google at a recent AI summit in Paris. Not in a big way. Not as a proclamation. As an analogy for his own behavior, consider how you might realize you haven’t watched cable TV in six months. “I realized, over the last couple of months, that I don’t use a lot of the services that I used to use,” he stated. The audience chuckled a little. Most likely, those whose businesses rely on search traffic did not.

The current state of search has the appearance of moving slowly at first, then all at once. The basic process of finding information online was the same for about thirty years: type some keywords, get a ranked list of links, click through, read, go back, and try again. It was accomplished by AltaVista. It was done by Ask Jeeves. Google built one of the most valuable companies in history by doing it better than anyone else. Over time, the interface underwent minor modifications, such as autocomplete, image tabs, and featured snippets, but the underlying logic remained constant. The engine pointed when you asked. You went and discovered the solution on your own.

Something structurally different is now challenging that model. Perplexity AI, ChatGPT’s search mode, and Google’s own Gemini-powered AI Mode are examples of tools that don’t just give you a list of links. They read a variety of sources, summarize their findings, and provide you with a well-written response that includes citations, options for further research, and the capacity to carry on the discussion in natural language. It feels more like asking a well-read coworker who has already read everything than using a library card catalog. 91% of participants in Applause’s 2024 Generative AI Survey report using chatbots for research, and 81% say they prefer them to traditional search engines for general questions. It is challenging to dismiss those figures as a niche shift due to their size.

It’s difficult to ignore how rapidly the product landscape has broken up. A few years ago, discussing a “AI search engine” would have elicited blank stares. These days, there are Perplexity, which has a devoted following among professionals and researchers seeking citation-backed answers across sources; Phind, which focuses on developers and pulls from technical documentation and code repositories; Consensus, which indexes academic literature and presents findings from peer-reviewed papers rather than general web content; and Felo AI, which manages multilingual research in ways that Google’s general search has never performed particularly well. Because Google was the only viable option, each of these is taking over territory that Google previously held by default. That is no longer the case, and the rate at which specialized tools are emerging indicates that this trend won’t change.

To be fair, Google isn’t standing still. In 2023, the company introduced its Search Generative Experience, which integrated AI-generated summaries straight into results pages. By the middle of 2025, AI Overviews—synthesized responses that appear before any conventional links—were appearing in about one in five Google searches. In just two years, Google Lens, its visual search product, has increased fourfold to handle 12 billion searches per month.

The company launched a full AI Mode in the UK in July 2025, delivering conversational responses instead of link lists using its Gemini model. It’s impressive to see how quickly Google is adapting; the company is not slow and has resources that no startup rival can match. However, there is a conflict in all of this adaptation: Google’s primary business strategy relies on advertising income linked to website clicks. An AI that provides a comprehensive response to your query directly on the results page is a product that undermines Google’s core business model. The business might be developing a product that rivals itself.

In late 2025, McKinsey calculated that AI-powered search could impact up to $750 billion in revenue by 2028, taking into consideration the companies that presently rely on search engine visibility to generate leads and sales. This number includes everything from media companies whose traffic is largely dependent on Google rankings to e-commerce retailers who optimize their product listings.

Not only does the transition from search to synthesis alter how people obtain information, but it also affects who is and is not seen. An AI response may use thirty sources but only clearly credit three, whereas a traditional search result yields ten links. The ramifications are substantial and still largely unknown for content publishers, SEO-dependent companies, and anyone who developed a digital strategy centered on being found on Google.

In the industry, there is a perception that people are actually giving up on keyword-based searching in general rather than Google specifically. This is because it requires you to translate your actual question into terms that a machine can understand, then piece together the answer from whatever fragments the links provide. A generation that was raised on it might not realize how strange it is until they use something that doesn’t require it. That translation step always felt a little awkward. It’s still genuinely unclear if the new tools are reliable, effective, and economically viable enough to completely replace a thirty-year-old habit on a large scale. However, the direction of travel is not.