The first edition of the Odyssiant Monthly AI Search Tracker. This month we look at how AI assistants have shifted from novelty to default starting point for B2B research - and what the data says about which brands are showing up.
Key takeaways
- ChatGPT, Perplexity and Gemini are now the first step in most B2B buying journeys
- Over 60% of shortlist queries return answers that never link to the brands mentioned
- Brands with structured proof assets (case studies, comparisons, third-party reviews) appear 3× more often
- Most companies still have zero visibility into what AI says about their products
- The gap between 'mentioned' and 'recommended' is where pipeline lives
AI is the new front door for B2B research
For most of the last two decades, "visibility" meant ranking on Google. If your page appeared on page one for the right keywords, you had a shot at the click, the visit, and eventually the conversion.
That model is not dead - but it is no longer the whole story. Today, a growing proportion of B2B buyers start their research by asking an AI assistant: "What are the best tools for X?", "How does Y compare to Z?", "What should I look for in a solution for this problem?"
The answers they receive are synthesised, authoritative-sounding, and often delivered without any outbound link. The buyer gets a shortlist - and may never visit your website at all.
What we're seeing across engines
This month we ran structured prompts across ChatGPT, Perplexity, Gemini and Claude, covering a range of B2B buying scenarios. Here's what stood out:
ChatGPT
ChatGPT continues to dominate as the default "first ask" for many business users. Its answers tend to be longer, more narrative-driven, and less citation-heavy. Brands that appear here are typically those with strong, well-structured content that has been indexed and absorbed into the training data. Real-time retrieval is improving but still inconsistent.
Perplexity
Perplexity leans harder on real-time web retrieval and citations. It regularly links to sources, making it the most "SEO-like" of the AI assistants. Brands with recent, authoritative content - particularly comparison pages, guides, and third-party reviews - tend to perform well here.
Gemini
Gemini's answers are typically concise and structured. It often provides bullet-point summaries and is less likely to elaborate unless prompted. Brands that appear in Gemini results tend to have strong structured data and well-formatted product pages.
Claude
Claude tends to give balanced, nuanced answers and is more willing to say "it depends." It frequently mentions brands that have clear differentiation messaging and transparent product information.
The visibility gap: mentioned vs. recommended
One of the most important distinctions we're tracking is the gap between being mentioned in an AI answer and being recommended. Many brands appear as part of a list - "Other options include X, Y, and Z" - without any endorsement or supporting evidence. Being mentioned is visibility. Being recommended is what drives shortlists.
The brands that consistently earn recommendations share a few things in common: they have proof (case studies, data, named customer examples), they have comparison content that positions them clearly, and they have content that directly addresses the buyer's question rather than talking about features in isolation.
What this means for your content strategy
If your content strategy is still built around keyword volume and search rankings, you're optimising for an incomplete picture. The question is no longer just "Do we rank?" - it's "When a buyer asks AI about our category, do we show up in the answer? And if so, are we recommended or just listed?"
The brands that are winning in AI answers are doing a few things differently:
- Building proof assets: Case studies, ROI evidence, named customer examples that AI can cite
- Creating comparison content: "X vs Y" pages, honest positioning against alternatives
- Answering real buyer questions: Content structured around the questions buyers actually ask, not just the keywords they type
- Maintaining freshness: Regular updates so content stays relevant to real-time retrieval engines like Perplexity
Next month
In next month's tracker, we'll be looking at how AI assistants handle "shortlist" queries - the moment a buyer moves from exploring to comparing specific solutions. We'll share data on citation patterns, which types of content earn links, and how answer structure varies across engines.