Let's start with the figure that should concern every B2B executive.
The 2X AI Visibility Index from April 2026, analysing 70 B2B companies, reveals that 96% of them are invisible in AI responses to discovery queries. Only 4.3% appear in responses to top-of-funnel queries — where shortlists are actually formed.
94% of B2B buyers now use AI in their purchasing process, and they cite generative AI as a more reliable information source than vendor websites, product experts, or sales representatives.
Place the two figures side by side: 94% of your buyers use AI to form an opinion — and 96% of B2B companies don't exist in those responses at the moment the opinion forms. This isn't a nuance. It's a structural disconnect between where the purchasing decision happens and where companies invest their visibility.
This article breaks down what that 4% does differently — and what our own observations add to this global data.
The critical distinction: discovery visibility vs confirmation visibility
The 96% invisibility figure requires an important clarification. The 95.7% of "invisible" companies appear primarily in queries where the buyer already knows their name — they're absent from the AI responses that shape shortlists upstream.
Appearing when someone searches your name is not the same as appearing when someone searches for a solution to a problem. The first is confirmation: the buyer already knows you, they're verifying. The second is discovery: the buyer is building their candidate list from scratch.
Almost all B2B companies are present in confirmation queries. Almost none are present in discovery queries. It's in this second type that shortlists are built — and therefore where commercial opportunities are won or lost before a salesperson ever enters the picture.
In our own tests, this distinction is verified systematically. A company can appear clearly when you type its name into ChatGPT, and be completely absent when you ask the recommendation question that corresponds exactly to its activity.
The four AI visibility profiles
Crossing data from the 2X AI Visibility Index with our field observations, four profiles emerge:
Profile 1 — Authority Leaders (4.3% of companies)
These are companies that appear consistently in discovery responses. They share three characteristics without exception:
A clear and consistent entity. Stable name, speciality formulated identically across all platforms, Schema.org correctly implemented. LLMs can identify them unambiguously.
Dense and recent third-party signals. Presence in recognised sector rankings, regular mentions in professional media, active review profiles with recent reviews. Domain traffic is the primary predictor of AI citations: sites with more than 1.16 million monthly visitors obtain an average of 6.4 citations per query, versus 2.4 for sites under 3,000 visitors. Audience builds authority, which builds visibility.
Content structured around precise use cases. Not generic educational content — pages that directly answer concrete purchasing scenarios. Pages with well-structured FAQ sections are 2.8 times more likely to be cited in AI responses than pages without FAQs.
Profile 2 — Strong Contenders (approximately 8% of companies)
These companies appear on queries corresponding precisely to their niche, but not on adjacent queries. They've built genuine sector visibility — without yet having the cross-cutting authority signals that would allow them to generalise.
This is strategically the most interesting profile: the actions to take are identified, the impact is measurable at 60-90 days. In our panel of 50 French SMEs, this profile corresponded to the most "activatable" companies — those for whom a diagnostic produces the most concrete results.
Profile 3 — Niche Players (approximately 12% of companies)
Strong visibility on a very narrow perimeter, near-absence on everything else. A law firm specialising in maritime law in Marseille that appears on all "maritime law lawyer Marseille" queries but on no adjacent queries. A HR SaaS publisher very visible on "leave management software for SMEs" but absent from all broader category queries.
This profile isn't a problem in itself — a well-defended niche can be commercially sufficient. It becomes a problem when the company wants to expand its addressable market without having built the authority signals that would make that move possible.
Profile 4 — Emerging & Lagging (75.7% of companies)
The dominant profile. These companies appear only on confirmation queries (their own name) or don't appear on anything at all. They've often invested in traditional SEO with good Google results — and discover that these investments don't automatically translate into AI visibility.
60% of AI citations come from URLs that don't feature in Google's organic top 20 for the same query. SEO history is not an advantage in itself in LLMs. What mattered yesterday doesn't matter the same way today.
What our tests add to global data
International studies measure trends on English-speaking panels often dominated by tech companies. On a French-speaking B2B panel of mid-size SMEs, we observe important nuances.
Fragmentation is even more pronounced in France and Belgium. The third-party reference sources for English-speaking LLMs (G2, Capterra, Clutch, Forbes, TechCrunch) have French-speaking equivalents that are less well indexed in global LLMs. A French SME can have an excellent reputation on e-marketing.fr or Le Journal du Net and remain near-invisible in ChatGPT, whose training remains predominantly English-speaking.
The language of the query changes results. In our tests, asking a query in French produced different results from the same query in English for the same companies. Companies that publish structured content in English on their areas of expertise appeared on French queries better than French-only competitors — because their entity signal was stronger in the global LLM index.
The Les Échos/Statista ranking is the French-speaking equivalent of G2. For consulting, recruitment and agency firms, this single source concentrates a disproportionate share of LLM citations on recommendation queries. Being listed or not creates a visibility gap that the best proprietary content cannot fill.
The signals that predict visibility: what the data shows
Crossing our observations with available data, five signals emerge as the most predictive of B2B AI response visibility:
1. Volume of recent third-party citations. This is the signal most correlated with visibility on Perplexity. Perplexity observes a 40% citation drop for content older than 30 days — the freshness of third-party mentions matters as much as their existence.
2. Presence in recognised sector rankings. On ChatGPT, whose responses rely more on training memory, rankings published by high-authority media (Les Échos, Gartner, Forbes) function as durable trust signals. A mention in the 2025 ranking will still be exploited in 2026 responses.
3. Density of reviews on verified third-party platforms. G2, Capterra, Clutch, Malt depending on sector. Sites implementing structured data and FAQ blocks recorded a 44% increase in AI search citations.
4. Entity consistency across platforms. Name, speciality, description: formulated identically across all online presences. This signal is underestimated and often neglected.
5. Structured FAQ content around discovery queries. Not the questions you want to answer — the questions your prospects ask an AI engine when they're looking for a provider like you.
The opportunity window — and why it's closing
The gap between the top quartile and bottom quartile in AI visibility terms grew from 32 points in Q4 2025 to 42 points in Q1 2026. The curve is accelerating.
Companies building their AI authority signals now have a structural advantage that competitors won't easily recover in 18 months. This isn't a projection — it's the pattern observed in SEO between 2010 and 2015, reproduced at higher speed.
The good news: unlike traditional SEO where authority builds over years, the signals that influence AI visibility can produce measurable effects in 4 to 12 weeks for engines like Perplexity. Perplexity is the most reactive LLM to an active strategy: a 0% citation rate can rise to 31% in 4 months after correcting missing signals.
What this changes for your strategy
If you're among the 95.7% of companies visible only on confirmation queries, the question isn't whether you need to act — it's where to start.
Our complete guide to AI visibility for B2B companies lays the methodological foundations. Our scoring tool analyses your current position across 6 dimensions in 5 minutes and identifies your priorities. For a complete analysis — visibility profile, benchmarking of your direct competitors on your target queries, prioritised action plan — our AI Diagnostic delivers a report within 5 business days.