How AI Decides What Answers To Share—And What That Means for PR & Marketing Teams

ChatGPT’s search traffic grew 44% month-over-month in late 2024, while Perplexity saw 71% growth during the same period. Additionally, a gartner report published at the beginning of January suggests the traffic from search engines to websites will drop by 25% by 2026. Backing this insane trend are recent figures which show that ChatGPT is now handling 2.63 billion monthly visits with an average session duration of 6 minutes – compared to Google’s average of 2 minutes. 

While there’s no denying that Google still claims the lion’s share of search volume, two points are becoming undeniably clear: 

1. That people are using LLMs to search and compare products and services online. 
2. For a period of time, there will be an unfair advantage for those that are early to adopt strategies that make them rank well in LLM search results.

Most companies though have no idea how AI uses sources to display information, and lack a clear understanding of how to influence what appears in these results. Even the tools tracking LLM rankings are extremely early in their development.

This leaves many sales and marketing teams operating with playbooks designed for traditional search landscapes.

Here’s what we’ve learned helping clients increase their AI Search visibility at mvpr – and subsequently what I think it’s important to know: 

Understand how AI displays search results – credibility factors

AI isn’t searching for results that match queries and ordering them first to last based on relevancy like Google does, it curates the sources that it believes will provide the most valuable answers and consolidates the information in them to display the information it thinks answers the question best – in one-shot. 

Like humans, when LLMs assess what information to display, they rely on a number of filters:

1. Relevance: AI matches queries using semantic similarity and keyword analysis. It understands context, not just exact word matches, evaluating how closely your content aligns with the user’s intent.
2. Credibility: AI prioritises trusted sources – like Bloomberg or SEC filings – over self-published content. Third-party validation carries significant weight in establishing your authority on a topic. 
3. Freshness: For time-sensitive topics, AI prefers recent content. A recent piece of research or product announcement will trump a big company profile done months ago.
4. Content Quality: AI filters out clickbait, duplicates, and shallow content. Detailed, insightful content consistently wins over marketing fluff and thin but keyword-rich content created purely for SEO.
5. Context Awareness: AI search systems track your conversation history to deliver personalised results, allowing follow-up questions without repeating background information and refining recommendations based on your specific situation and previously expressed interests.
6. Diversity –  AI prefers to present multiple viewpoints rather than one answer, drawing from academic, journalistic, commercial, and user-generated content to help users form informed opinions while consciously working to avoid information bubbles.

Together, these filters create a more sophisticated evaluation system than traditional search, requiring a less traditional approach to visibility.

Messaging Consistency for AI: Why consolidated AI answers mean you need to take messaging consistency seriously

AI search takes a ruthless approach to redundant or confusing information – it filters it out. 

LLMs champion clear, information-rich, insightful content – and devalues repetitive, low-quality sources. When asked to recommend products or companies, AI makes its analysis based on a comprehensive evaluation framework that combines user context, product quality, and alignment with needs:

It matches intent: AI first interprets what kind of recommendation you want: best overall, best for a specific use case, or best within constraints like budget or geography.

Gathers User Context: AI incorporates what it knows about your role, industry, or past queries to tailor recommendations to your specific needs.

Ensures Feature Fit & Use Case Alignment: AI reads product pages and prioritises those with core features that directly address the problem in your query.

Prioritises Source Quality: AI relies on recent, credible sources and third-party analysis that gives evidence that the product works in practice, not just in theory. Like humans it’s more likely to trust someone else’s opinion of your skill, than your own. 

Differentiation & Market Signals: AI looks for standout capabilities, unique positioning, and signs of traction like funding, customer reviews and user adoption.

Tradeoff Clarity: AI highlights pros and cons so you understand where a product excels and where it might fall short of your requirements. 

Bias Avoidance & Balance: AI avoids commercial bias, aims for objectivity, and to include both established leaders and rising challengers.

If your website describes you one way while press mentions position you in another, the AI search won’t be confident that what you do meets the requirements of the searcher, and it will filter you out. 

How to rank well in AI search results

To ensure your company performs well in AI search results, you need a coordinated approach to targeting the sources that are prioritised:

• Get your PR right and build credibility with industry media and respected publications that are well-ranked by LLMs through expert commentary and balanced thought-leadership
• Focus on brand mentions, the more well-recognised the news site, the better. Backlinks are less important for LLMs but naturally still valuable for traditional SEO. 
• Ensure your website blogs are detail-rich and provide balanced perspectives on your company, customer challenges, or industry
• Use accurate, verifiable data in your content rather than vague claims
• Encourage other experts to reference your articles or work in their own content
• Apply for awards, speaking slots at relevant conferences, podcasts, and ensure review sites have accurate information
• Structure your website clearly with consistent messaging across all landing pages and make sure your About section is well-written and describes what you do
• Build brand presence across multiple platforms: Reddit, LinkedIn, YouTube, Substack, Stack Overflow, Medium, etc.
• Maintain consistent messaging across each channel so that it re-confirms what you do rather than confuses the model

PR for AI Search: focus on brand mentions that combine niche-specific and high authority publications

As models become more sophisticated at evaluating and consolidating information, and as more companies get wise about AI search, the bar will be raised for what information is included in the sources. 

Assuming AI models continue to prioritize information based on the criteria above, as search volumes grow and more users turn to AI for information discovery, strategic and consistent PR is going to become more important than ever. Focus on building a coherent, consistent presence, backed by third-party and earned validation.

This is going to disproportionately reward companies that can provide evidence to support their expertise and are able to communicate well about it. It will punish companies with fantastic ads but terrible user reviews, or keyword rich content that is thin on substance – which we would probably all agree is a good thing. 

AI Search, unlike traditional SEO, is significantly harder to manipulate through technical tricks. It rewards what successful businesses have known all along – create an incredible experience for your customers, and authentically communicate about that. 

AI search feels like the search we should have had all along, and right now for those first to the garden – there is a lot of low hanging fruit.