ChatGPT is not just answering questions; it is searching like a power user. The model now performs live web searches in almost one-third of prompts, and those searches are longer, more detailed and more commercially focused than anything seen in traditional search behaviour. With more than 800 million active users, according to OpenAI’s latest figures, ChatGPT is no longer a niche productivity tool. It has become one of the largest search interfaces in the world, and its habits are starting to matter.

New research examining over 8,500 ChatGPT interactions has revealed how the model structures its searches, what types of content it seeks out and where opportunities exist for brands to appear in its results. Conducted by analytics firm Nectiv, the study sheds light on how conversational AI is reshaping search intent and what that means for marketers competing for visibility inside large language models.

ChatGPT’s Search Mechanics

In nearly a third of all prompts analysed, ChatGPT initiated a web search. Each time it did, it performed an average of just over two queries, with some prompts triggering as many as four. The typical ChatGPT search was 5.48 words long, around 60% longer than a standard Google query. More than three-quarters of searches contained five words or more, signalling a marked shift towards longer, more specific phrasing.

These extended queries reflect how people naturally express intent in conversation rather than how they typically type into a search box. ChatGPT’s searches tend to resemble research briefs rather than keywords, such as “best boutique hotels near Lisbon old town reviews” or “compare Razer Phone 2 vs Razer Phone 1 user feedback.”

The most common modifiers in ChatGPT searches were “reviews,” “features,” “comparison” and “2025.” The inclusion of the current year is particularly telling, revealing how the model prioritises freshness and relevance when sourcing external information.

What ChatGPT Searches For

The dataset showed significant variation in how ChatGPT searches across different industries. Local intent dominated, appearing in 59% of cases where a search occurred. Commerce followed at 41%. At the other end of the spectrum, categories such as credit cards and fashion saw the lowest frequency of external search, at 18% and 19% respectively.

Local queries triggered more searches than any other category but produced fewer total fan-outs, averaging 1.67 searches per prompt. ChatGPT tends to find what it needs quickly when intent is location-based. In contrast, complex sectors such as jobs, careers and software generated deeper search patterns, with an average of nearly three queries per prompt. These verticals often require layered data such as pricing, features or industry comparisons, prompting ChatGPT to explore multiple information sources before forming an answer.

Across the dataset, the longest queries appeared in real estate and credit cards, typically exceeding six words. These often included comparative and financial language, reflecting the complexity of decision-making in those categories.

The Shift To Conversational Intent

ChatGPT’s approach marks a sharp departure from how search has historically worked. Instead of short, keyword-led commands, the model relies on full-sentence, context-rich queries that mimic how humans think and talk. Each ChatGPT prompt can generate several related searches, a process known as “fan-out,” allowing the model to cross-reference results and synthesise the best response.

This structure rewards content that is comprehensive and semantically rich rather than content built purely for keywords. A single well-optimised page covering related questions and comparisons is more valuable to ChatGPT than several thin pages targeting isolated phrases.

For marketers, the implication is clear: search visibility is now governed as much by conversational relevance as by keyword ranking. Pages that reflect the way users naturally articulate questions and decisions stand a far greater chance of being surfaced in AI-assisted discovery.

Optimising For ChatGPT Discovery

Traditional SEO principles still matter, but the Nectiv data shows that the levers of optimisation are shifting. Brands that want to remain visible within ChatGPT’s search layer should consider the following strategies.

1. Build for specificity. Long-tail, descriptive phrases mirror the way ChatGPT searches. Content should capture detailed attributes such as model names, release years, pricing tiers and locations.

2. Emphasise reviews and comparisons. Content that evaluates or benchmarks products performs well, reflecting ChatGPT’s preference for sources that aid decision-making.

3. Update regularly. The frequency of “2025” in searches highlights the model’s emphasis on freshness. Regular updates to content titles, metadata and examples can help signal recency.

4. Use structured and scannable formats. Clear formatting, bullet points and schema markup make it easier for ChatGPT to parse information efficiently.

5. Cover topic clusters, not isolated keywords. Because ChatGPT performs multiple connected searches, comprehensive content that anticipates related queries is more likely to be referenced.

Beyond ChatGPT

While ChatGPT remains the most visible player, it is not the only large language model reshaping online discovery. Anthropic’s Claude and Perplexity AI’s conversational search tools are adopting similar grounding techniques, blending LLM reasoning with real-time web data to generate contextual, source-aware answers.

Google is moving in the same direction through AI Overviews and the rollout of AI Mode, which combine traditional search listings with generative summaries and conversational follow-ups. These features are transforming Google from a directory of links into an interpretive engine that can answer, advise and predict in real time.

For marketers, optimisation now extends beyond any single platform. The next frontier of visibility will be LLM discoverability: ensuring that brand content is legible, credible and accessible to any AI system drawing from the open web.

This shift requires marketing teams to think less about search engines and more about AI readers. Models like ChatGPT and Google’s evolving search products no longer simply index pages. They interpret them, summarise them and present them to users as authoritative answers. The opportunity lies in ensuring that your brand’s insights, data and perspective are the ones these systems choose to surface.

Implications For Marketers

The rise of AI-driven search introduces a new dimension to visibility. Rather than relying on a single query-result dynamic, brands must now consider how AI systems retrieve, verify and synthesise information.

This evolution blurs the boundaries between SEO, content strategy and AI optimisation. It also amplifies the value of structured, high-quality data. For sectors such as B2B software, travel, real estate and local services, where ChatGPT search activity is already high, the opportunity is to create authoritative content that directly supports complex intent.

Longer, review-led and comparative searches indicate that users are often close to purchase decisions. If ChatGPT and Google’s AI systems are grounding their responses in web content at this stage, the ability to appear in those summaries could shape consumer choice well before a site visit ever occurs.

The Next Stage Of Search

ChatGPT’s growing search behaviour and Google’s AI integration signal a convergence between traditional SEO and AI discovery. The model’s reliance on longer, context-driven queries represents a fundamental change in how digital information is found and evaluated.

For marketers, the challenge is to adapt faster than the systems evolve. Optimising for AI visibility means thinking like an algorithmic reader: precise, structured, relevant and current.

Search optimisation is no longer about ranking highest in a list of links. It is about being selected by the systems now deciding what the world sees first.