AI Visibility Barometer
2026 Edition

Why AI visibility matters now

In 2026, LLM-powered search engines have become a primary discovery channel for B2B buyers. When a RSSI, DSI, or IT manager asks ChatGPT "what is the best cybersecurity solution for a French SMB subject to NIS2", the answer they receive directly shapes vendor shortlists — before a single web page is visited.

Unlike traditional search engine rankings, LLM citations are not disclosed, not bid-able, and not tracked by conventional analytics. Most companies have no idea whether they appear in these responses — or how frequently, and at what rank.

The AI Visibility Barometer was created to make this measurable.

The visibility gap is structural

The 2026 pilot reveals a highly concentrated distribution of AI visibility in the French B2B cybersecurity sector. A small number of companies dominate LLM responses across all four tested models, while the vast majority — 73% — are effectively absent.

This is not primarily a brand size effect. Several well-funded companies score below 20, while some smaller, content-focused operators achieve scores above 60. The data suggests that AI visibility is driven by content signals, structured data, and citation corroboration — not by budget or market share alone.

Not all LLMs behave alike

Citation patterns differ significantly across the four tested models. Perplexity returns the broadest set of distinct company citations per prompt, likely reflecting its real-time web retrieval architecture. Claude and ChatGPT show stronger concentration — a smaller set of companies dominate their responses more consistently. Gemini sits between the two patterns.

These differences have strategic implications: a company invisible on ChatGPT may have meaningful visibility on Perplexity, and vice versa. Per-LLM score breakdowns are available in the full dataset.

Signal patterns in high-scoring companies

Companies in the top quartile (score 60+) share common characteristics:

• Consistent, expert-authored content on category topics (GRC, EDR, NIS2, SOC)
• FAQPage and structured data markup across key pages
• Independent third-party citations (press, analyst reports, comparison sites)
• Strong entity recognition — same company name used consistently across all sources
• Category-specific landing pages matching the prompt language of buyers

Full analysis and recommendations are available in the PDF report.