AI Search Has Made Search Strategy a Moving Target

You know the tools by now: ChatGPT, Gemini, Perplexity, Google’s AI Overviews. Odds are you have used at least one of them this week, and so have your buyers.

A prospect researching your category may still open Google, but increasingly they are asking an AI assistant to compare vendors, explain a market, or recommend a shortlist before they ever visit your website. Gartner projects traditional search engine volume will fall 25% by 2026 as users shift to AI chatbots and agents.

Discovery has become more conversational, more fragmented, and increasingly shaped by systems that synthesize information instead of simply returning a list of links.

The Rules Refuse to Sit Still

The temptation right now is to look for a repeatable formula. Which websites do AI platforms prefer? How should an article be structured? What earns a ChatGPT citation?

The problem is that the answer keeps changing.

Reddit is one example. Marketers moved toward the platform because its archive of firsthand product discussions made it an important source for AI systems. Then in August, Axios reported Promptwatch data showing a sharp decline in Reddit’s share of citations inside ChatGPT Search, with no clear explanation.

That does not mean Reddit suddenly stopped mattering. It means reorganizing your content strategy around one source or one observed citation pattern is risky. You can execute the tactic perfectly and still end up optimizing for yesterday’s behavior.

Click behavior is shifting too. An Ahrefs study of roughly 300,000 keywords measured a 34% drop in clickthrough rate for top-ranking pages when an AI Overview appeared. Google disputes that framing and argues AI-era clicks can be higher quality.

Whoever wins that argument, the larger point is the same: the ground keeps moving underneath any strategy tied too closely to one metric or one platform.

Your Content Now Competes Inside the Answer

A client asked me recently whether we should start writing articles “for ChatGPT.”

I understand the question, but I think it assumes AI search works like traditional search: study the ranking factors, then produce content to match them.

That is not really the opportunity.

An AI system composes the answer itself, which means generic explanations are the easiest thing for it to produce. For years, content marketing rewarded volume. Publish more articles, cover more keywords, feed the index, and capture the traffic that came back.

Language models change that math.

An article that simply restates what a hundred other articles have already said brings very little new information to the system. The model can already explain the category.

What it does not have is what only you know: your data, your customers’ measured outcomes, your firsthand experience, and the evidence your company has collected.

That is the material worth building around.

Instead of asking, “How do we write for ChatGPT?” the better question is:

What do we know that the model cannot get anywhere else?

The harder part is what happens to traffic. As AI systems take over more of the explaining, visits may fall even for brands whose influence is growing. Buyers can arrive at your website later in the process, sometimes with a shortlist already formed through a conversation you never saw.

That changes the role of the website. It starts to function less like a storefront and more like a source document.

And that is why putting your strongest proof behind a form deserves another look.

If your best customer outcomes, original data, or strongest evidence cannot be accessed, you may be hiding the exact information you want buyers and AI systems to associate with your company.

Run Visibility Like a Portfolio

Measurement has to change along with everything else.

A ranking is a position you can check. An AI answer is closer to a snapshot of how the system currently understands your market and your company.

That means visibility is better sampled than obsessively tracked.

We have found it more useful to keep a fixed set of questions buyers actually ask, run them regularly across the platforms they use, and record whether the brand appears, how it is described, and whose evidence the answer relies on.

Daily fluctuation is noise. What matters is whether the overall picture is changing, and in which direction.

That volatility also creates a distribution problem. If one platform’s citation influence can rise or fall quickly, relying too heavily on any one channel becomes a risk.

Owned content, earned media, executive voices, community discussion, and original research all have a role. The goal is not to predict which channel will matter forever. It is to make sure your brand is supported by enough credible evidence across the market that you are not dependent on one source staying influential.

The Reframe

I suspect GEO will eventually stop feeling like its own discipline.

Optimization tactics will change because the underlying systems will change. What will remain is the work of being worth referencing: evidence, consistency, original information, credible customer outcomes, and named experts standing behind the work.

PR has operated on a version of this logic for a long time. The intermediary used to be an editor deciding what an audience should hear.

Now the intermediary may be a model.

The job has not changed as much as it seems.

Where the platforms go next, nobody honestly knows.

Build for the constant instead.

Machines, like the buyers they serve, quote the sources they trust.

‍