The Big Four Are Telling You the Web Has Changed

Because they are describing the product they are building, not making predictions. When a single company reframes search you can call it positioning. When the four companies that between them shape almost all of search reframe it the same way, within the same short window, it is a change in the ground beneath your marketing.

Here is what each of them is actually saying, in plain terms, and where they agree.

Google now describes its own search as an AI experience. Its AI Overviews sit above the links on a large share of results, and its AI Mode goes further, breaking a question into many related searches and synthesising one answer. Google has said AI Mode reached around a billion people, a figure it reports itself. The mechanism behind it is query fan-out, which we explain separately.

OpenAI built search into ChatGPT around cited answers rather than a ranked list, and struck content partnerships with publishers so those answers can draw on trusted sources. The framing is telling: the product is an assistant that answers and attributes, not a directory you browse.

Microsoft has said for some time that being visible is no longer only about blue links, and has folded AI answers through Copilot into the search experience. Its guidance to site owners has shifted accordingly, toward being a source an AI system can understand and cite.

Perplexity does not position itself as a search engine at all. It calls itself an answer engine, presents a synthesised response with numbered sources by default, and runs a publisher programme that shares revenue when those sources are used. The entire product is built on the citation.

On the essentials, completely. The table lays their positions side by side.

Company

What their words and products say

Google

Search is now an AI experience; AI Overviews and AI Mode answer above the links, using query fan-out

OpenAI

Search means a cited answer inside ChatGPT, backed by publisher partnerships

Microsoft

Visibility is more than blue links; Copilot brings AI answers into search

Perplexity

Not a search engine but an answer engine, with sources cited by default and revenue shared with publishers

On the direction, yes. The Princeton-led study that named generative engine optimisation measured what lifts a brand inside AI answers and found statistics, quotations and cited sources were the strongest levers, worth up to about a 40% gain. That is independent confirmation that the cited answer is now the surface that matters.

But there is a warning alongside the trend. Independent analysis of AI search has repeatedly found accuracy problems.

Columbia’s Tow Center found that more than 60% of AI-search responses it tested contained factual errors.

Read together, the research says two things at once. The cited answer is where visibility now lives, and the engines get things wrong often enough that being cited is only half the job. Being cited accurately is the other half, and that is squarely within your control.

It changes the question you ask. Not “where do we rank?” but “when an answer engine describes our category, are we named, and is it right about us?” The way to influence that answer is to give the engines clear, well-structured, well-sourced material about your brand, which is the work of GEO. If you are new to it, start with how to do GEO.

One company reframing search is a strategy. Four doing it in unison is a shift. The brands that adjust now, structuring themselves to be cited and cited accurately, will be the sources these engines lean on tomorrow.