LLM Optimisation
GPT-5 Changes the Game for AI Search – Here’s What Marketers Need to Know
OpenAI has just rolled out GPT-5, and while most of the headlines are about its speed and smarts, there’s a change under the hood that’s a big deal for anyone in marketing, especially if you care about product discovery and brand visibility.
21 August 2025
5 min read
One Model to Rule Them All
With GPT-5, there’s no more picking between 4.0, o3, o3-mini, and other versions. You ask a question, and the system decides whether to give you a fast answer, run deep reasoning, or even fire up a live web search. This happens automatically.
For marketers, this means shopping and product research queries are far more likely to use reasoning models, the heavy-duty versions that gather and process huge amounts of information before responding.
Reasoning Models Are Long-Tail Hunters
Over the past six months, we’ve seen how reasoning models behave when used for shopping: they can pull from dozens (sometimes 50–100) different sources to build a single answer.
This breaks the old “page one” mindset. Rankings on Google’s first page won’t mean much when GPT-5 is happily gathering content from what would have been page 10 or even page 20. The AI will synthesise from a wide range of sources, much like a human doing serious research.
The Implication: Breadth Beats Position
Because no single page is the “gateway” anymore, brands will need to optimise across a broader set of pages and formats. Breadth, volume, and velocity of relevant content will become the winning strategy, not just owning a few high-ranking pages.
Mass Adoption Without the Friction
Until now, using a reasoning model was a choice you had to make in the UI. Now, GPT-5 picks it for you, meaning more people will start using it without realising. The result? Far richer, deeper, and more accurate product research experiences will become the default.
For example, we recently used o3 (a reasoning model) to research running shoes. It spent over six minutes reviewing 48 different websites before giving a shortlist. That’s the level of diligence people will start getting automatically.
AI Visibility Just Got 100x More Interesting
With GPT-5, AI-powered search is no longer an optional extra, it’s baked in. The playing field for brand visibility has been reshaped overnight. Understanding how these models find and compile information is now mission-critical for marketers.
The bottom line: search as we knew it has changed. The question now is whether your brand is ready to be found in this new AI-driven reality.
With GPT-5, there’s no more picking between 4.0, o3, o3-mini, and other versions. You ask a question, and the system decides whether to give you a fast answer, run deep reasoning, or even fire up a live web search. This happens automatically.
For marketers, this means shopping and product research queries are far more likely to use reasoning models, the heavy-duty versions that gather and process huge amounts of information before responding.
Reasoning Models Are Long-Tail Hunters
Over the past six months, we’ve seen how reasoning models behave when used for shopping: they can pull from dozens (sometimes 50–100) different sources to build a single answer.
This breaks the old “page one” mindset. Rankings on Google’s first page won’t mean much when GPT-5 is happily gathering content from what would have been page 10 or even page 20. The AI will synthesise from a wide range of sources, much like a human doing serious research.
The Implication: Breadth Beats Position
Because no single page is the “gateway” anymore, brands will need to optimise across a broader set of pages and formats. Breadth, volume, and velocity of relevant content will become the winning strategy, not just owning a few high-ranking pages.
Mass Adoption Without the Friction
Until now, using a reasoning model was a choice you had to make in the UI. Now, GPT-5 picks it for you, meaning more people will start using it without realising. The result? Far richer, deeper, and more accurate product research experiences will become the default.
For example, we recently used o3 (a reasoning model) to research running shoes. It spent over six minutes reviewing 48 different websites before giving a shortlist. That’s the level of diligence people will start getting automatically.
AI Visibility Just Got 100x More Interesting
With GPT-5, AI-powered search is no longer an optional extra, it’s baked in. The playing field for brand visibility has been reshaped overnight. Understanding how these models find and compile information is now mission-critical for marketers.
The bottom line: search as we knew it has changed. The question now is whether your brand is ready to be found in this new AI-driven reality.