Business visibility after May 2026: what changes, what remains

Note revised on 25 May 2026. Article originally published in September 2024 — full rewrite following the Google Search Central publication of 15 May 2026. This note absorbs the content of pieces previously titled "How to be cited by ChatGPT: 7 actionable strategies", "Regaining control over what AI says about your company", "Generative reputation: when AI becomes your audience" and "AI visibility audit: the complete guide for executives", de-indexed and redirected to this page.

In May 2026, two events shifted the grammar of business visibility without overturning it. On 15 May, Google Search Central published its page Optimizing your website for generative AI features on Google Search, filed among the Search Central fundamentals alongside the SEO Starter Guide^[1]. Four days later, Google I/O 2026 unveiled the extension of AI Mode, the continuity between AI Overviews and AI Mode, multimodal search, and the first search agents able to execute tasks on the user's behalf^[2].

Taken together, these two events change what the word visibility covers, without changing what produces it.

What changes

A company's visibility in Google is no longer merely a position in a list of links. It is also its presence or absence in a generated response, its likelihood of being cited, contextualised or recommended in a synthesis that the user may not follow with any click.

A correctly indexed but editorially weak company may thus find itself absent from generated responses on its most strategic queries, without having lost a single organic position. Conversely, a company with a clear editorial voice and rigorous sourcing may be cited in generated responses without appearing in the very first ranking positions. The selection criterion for citation is not exactly the selection criterion for ranking. The two still correlate, but imperfectly.

This dissociation alters the way B2B buyers and individuals arbitrate their decisions. On complex queries — comparing several providers, understanding a topic, preparing a choice — the generated response now often precedes the consultation of links. The user reads the synthesis, retains the sources cited in it, and sometimes has no need to scroll further down.

What does not change

For two years, the digital consulting market sold a new discipline supposedly called GEO, allegedly distinct from SEO through its own principles, specific tools and novel methods. This presentation is now at odds with Google's official doctrine.

The Search Central page of 15 May 2026 sets out a structuring position: generative features rely on the central ranking and quality systems that already govern classical search. They add specific generation techniques — notably retrieval-augmented generation — but the eligibility bedrock of content remains familiar: helpful and non-derivative content, technical crawlability, transparent sources, authority signals built over time. No new acronym, no new profession, no skill to acquire separately.

What changes is not the nature of the fundamentals. It is the severity with which they are applied. In a list of ten links, mediocre content could appear in seventh position and capture residual traffic. In a generated response that cites a few sources, mediocre content no longer appears. The bar has been raised. What used to pass no longer passes.

The measurement that was missing

This raising of the threshold produces a practical situation: traditional visibility indicators — organic position, SEO traffic, ranking against competitors — remain valid but no longer suffice to describe a company's actual situation. A marketing department that tracks only these indicators looks at part of its visibility. That part remains important. It no longer exhausts the question.

A measurement must therefore be added. A measurement that applies to the synthesis regime, that runs on the generative models where this synthesis takes place, and that produces an indicator a management team can track over time.

MCVA Consulting SA has stabilised this measurement under the Score GEO™ name. The term calls for a precision that grounds the firm's doctrine: GEO as a sellable discipline left the vocabulary on 15 May 2026; the Score GEO™ as a proprietary instrument for measuring AI citability, on the other hand, remains. We do not sell a new skill. We measure a real visibility.

This measurement aggregates observations conducted on five generative models — ChatGPT, Claude, Perplexity, Gemini, Mistral — from business-anchored query baskets, according to a grid of seven themes: citational presence, editorial prominence, citation context, factual accuracy, source anchoring, cross-model coherence, temporal stability. It produces an overall score out of 100, accompanied by a detailed thematic profile that clarifies the gaps to be addressed.

Going further

The Score GEO™ method, its basket construction, its evaluation grid and its operational mode of use are set out in detail in the Cahier MCVA n°1 — Measuring AI citability after May 2026, the first volume of the Cahiers MCVA collection. The Cahier also offers a reading by typology of Swiss companies and a roadmap for marketing departments that wish to integrate this new measurement into their existing apparatus.

What is taking shape is not a rupture to prepare for. It is a measurement to add.

Sources

[1] Google Search Central, Optimizing your website for generative AI features on Google Search, 15 May 2026. developers.google.com/search/docs/fundamentals/ai-optimization-guide [↩]

[2] Google, Google Search's I/O 2026 updates: AI agents and more, 19 May 2026. blog.google/products-and-platforms/products/search/search-io-2026/ [↩]

Jérôme Deshaie is CEO of MCVA Consulting SA, a Swiss firm specialising in strategic consulting on artificial intelligence, based in Valais.