Generative AI has transformed how users ‘search’, as people prefer conversational, synthesized answers instead of static lists of links. And similar to how Google became synonymous with search, ChatGPT has become synonymous with AI search.

Developed by OpenAI, ChatGPT is one of the many generative AI chatbots that have accelerated the ongoing AI boom. However, none has been as successful as ChatGPT. Despite recent gains in market share by competitors, it still retains the majority share (53%) as the most used AI engine worldwide.

These results reflect our study findings. Agility’s ‘Closing the AI Visibility Gap’ report asked 112 PR professionals to name their top three priority AI answer engines:

• ChatGPT led at 85%
• Google AI Overviews (Gemini) at followed at 61%
• Microsoft Copilot came in third, with 35% of respondents

Almost every team is monitoring ChatGPT, which makes it the natural starting point for your AI visibility. Let’s delve deeper into how ChatGPT works, how it cites, and the takeaways that can guide your AI visibility strategy.

What our research revealed about ChatGPT behaviour

As part of our study, we submitted 6,120 prompts to ChatGPT across 48 industry categories. This helped us understand how the platform functions. Here are some key insights from our research:

ChatGPT has the smallest citation pool

ChatGPT averaged 3.6 citations per answer, the lowest among the four AI engines we examined. That’s about 4 fewer citations per prompt than the second-lowest, Perplexity, at 7.7. This suggests the platform relies heavily on its training data and, when it conducts a live search, draws on a highly targeted list of sources.

In contrast, ChatGPT cites 6.3 brands per answer, similar to Perplexity (6.8) and Claude (6.2). Listing 6 brands per prompt while drawing on just 3.6 citations highlights a list that is harder to break into. A brand must consistently appear across citations to make it to the list.

It has the highest number of answers without citations

ChatGPT not only cites the fewest sources; it also has the highest rate of no-citation responses. Our research shows that all other platforms cite at least one source in at least 99.3% of their responses. However, citations appear in only 94.5% of ChatGPT answers.

The nearly 5% gap reveals a key insight into ChatGPT behaviour. It is the one engine that will still name brands from model knowledge alone rather than from retrieved sources. Industry research backs that up: a study reveals ChatGPT conducted a live search for roughly a third of prompts asked. The rest were answered through training data and model memory knowledge.

For PR teams, this shows that some of ChatGPT’s brand recommendations are not influenced by current brand representation across AI engines. These recommendations may be shaped by how your brand has performed historically. With new models and updates coming regularly, it is important to maintain a consistent presence across your industry over the long term to influence training data recommendations.

Mainstream press carries more weight on ChatGPT than anywhere else

This is the clearest behavioural difference in our data. Major media accounts for 6.6% of ChatGPT’s citations, roughly double Claude (3.3%) and Gemini (3.1%), and more than four times Perplexity (1.4%).

The gap widens sharply by sector. In Beauty and Hair Care, major media accounts for 36.1% of ChatGPT’s citations. It is the only engine-industry pairing in the entire study where corporate content is not the top citation source, underscoring the influence of major industry publications.

Similarly, major media citations in Travel reach 16.2% and in Finance 12.9%. At the other end, social (0.3%) and forum (0.3%) citations are close to non-existent, the lowest of any engine tested.

This ties in with our earlier findings of the platform focusing on a select few citation sources. PR teams focusing on ChatGPT visibility need to integrate major media publications and collaborations as a cornerstone of their outreach strategy since community content and social conversation do not influence ChatGPT,

Performance drives inclusion more than on any other engine

80.88% of ChatGPT’s brand recommendations are driven by performance, based on how well a product works and what it delivers. Performance is a common driver of brand recommendations across all four engines, but it has the highest share in ChatGPT.

Across the four platforms, reputation influences the fewest brand recommendations in ChatGPT answers, at just 12.1%. Price (3.1%), trust (1.5%), and popularity (0.5%) barely register in the mix.

Awareness-led messaging has less traction on ChatGPT than anywhere else in the study. Brands are surfaced for demonstrated capability, not for being well known.

What this means for PR teams

AI visibility is not a one-size-fits-all landscape, and optimizing for brand visibility on ChatGPT requires PR teams to understand the distinct patterns and behaviours it displays.

Here are some practical takeaways for PR teams prioritizing ChatGPT:

• Consistent visibility is key. ChatGPT cites the fewest sources and regularly draws on training data, implying the need for consistent visibility and niche authority for brands to make the list.
• Mainstream media placements weigh higher here than for other engines. A major magazine or industry publication feature is more likely to get cited on ChatGPT than it is on Perplexity or Gemini. If your audience uses ChatGPT, tier-one earned media is a key visibility factor, not a vanity metric.
• Lead with evidence. While this is true across AI platforms, it is particularly helpful in getting cited on ChatGPT. Benchmarks, measurable results, outcomes, and comparative data are what drive brand recommendations.
• Provide key information upfront. Research shows that ChatGPT cites heavily from the top of the page, with the bottom half acquiring just 20% of citations. You have a higher chance of getting cited and getting your brand listed if you front-load critical information to earn the reader’s (and crawlers’) trust.

The AI answer engine your audience uses determines your visibility strategy. As the landscape evolves, PR teams need to regularly monitor where and how their brands appear across AI to understand user behaviour and adjust their approach.

Advanced PR tools need to integrate dedicated AI visibility measurement capabilities to empower communication professionals to track AI visibility, audit reputation, identify influential sources, and guide an informed AI visibility strategy.

 

That is exactly the challenge Agility’s Visibility Intelligence solves. Monitor your brand’s AI visibility, audit reputation, and identify what’s driving influence across ChatGPT, Claude, Perplexity, and Gemini directly within the platform.

Book a demo now to see how it can transform your AI monitoring strategy.