What you need to know from a communications perspective

As AI continues to reshape the communications landscape, marcoms directors should be continually assessing how they approach search visibility, content authenticity, and audience trust.

Here are my insights from a recent PR Moment AI webinar revealing the critical shifts in how people are using AI and the impact that is having on how brands should show up.

AI is changing search behavior

With 60% of searches now zero-click users are increasingly relying on AI summaries rather than browsing links that would have been thrown up by searches in the past. This shift from critical thinking to anchor bias means your brand’s visibility depends more on how AI interprets and presents your content.

Currently, roughly 90% of AI visibility is driven by citations from earned media. The LLM’s that feed into AI search results have a heavy bias towards trusted heritage publications (think Forbes, TechRadarPro, Good Housekeeping) as well as social platforms like Reddit and Utube. Therefore, to stay ahead brands should:

• Engineer multiple conversations across multiple platforms.
• Activate communities and advocates to influence peer-to-peer platforms.
• Be aware that some of the social platforms, while great for GEO, attract bad actors who spread misinformation.
• Identify any misinformation and correct it at source.

GEO is becoming king

GEO results are influencing user decisions, so it’s helpful to understand how AI ranks sources and when it turns to the web for information rather than using its training data. Claude’s leaked framework shows it uses ‘do not search but offer’ when it thinks it knows the answer, but recognises there might be newer information available on a topic. But it requires the user to ask it to check. ChatGPT source code analysis shows how its ‘re-ranker’ model will retrieve larger source sets and re-rank them as it seems most relevant and authoritative. It also has a ‘freshness scoring profile’ which favours new over older material.

Can we trust AI?

We also need to consider whether we can trust the information provided. From recent events, like the death of Charlie Kirk, where it was reported he was alive by AI after his death, we know that AI can’t handle breaking news well, it only demonstrates 85-90% accuracy for this type of news.

When LLM’s turn to the web for answers it can lead to hallucinations and misinformation, that is why earned media and PR are still so important for brands.

Interestingly, Brain Labs found for news queries Google still reveals the most applicable links. And AI Mode (Google/Gemini) and ChatGPT rarely cite from the same sources. AI Mode has a 28% overlap with Google, and ChatGPT has a 23% overlap.

So, can we trust these sources?

Google rarely cites from a low domain, so is the most trustworthy, whereas ChatGPT is four times more likely than Google to cite from a low quality domain. And AI Mode is three times more likely to do so than traditional Google results. Marcoms directors should:

• Use citation analysis to understand how their brand is represented.
• Defend against misinformation by feeding verified, high-quality content into the ecosystem.

SEO Isn’t Dead

Traditional SEO still matters. In fact, 25% of AI-generated search results rely on Google, showing that Google’s search index remains influential. However, AI platforms like ChatGPT and AI Mode (using Google’s AI Gemini) are changing how content is surfaced. Potentially LLM’s favour freshness over evergreen content, but marcoms directors should test this continually as it’s constantly changing. In essence, brands need to:

• Update content regularly to stay visible.
• Test how their site ranks in AI-powered search environments.
• Optimise for both traditional and AI-driven search engines.

Separate Generation from Verification

AI can produce polished but misleading content. As Felix Danczak noted, “Machines lie to us – polish is not accuracy.” He recommended treating AI like an intern who has read the entire internet – they are exceptionally useful for draft content, but a human must take responsibility for the content, verify it, and stand by the content. Marketing Directors should ensure:

• Human verification of all AI-generated content.
• Clear authorship and sourcing.
• Train staff in cross-examination and fact-checking.

Marcoms directors should embrace AI’s efficiency but be conscious of defending their brand’s authenticity. The winners will be those brands that combine technological savvy with human judgment, ensuring their content is not just seen but trusted.