Perplexity and Google after May 2026: complementary search regimes
Note revised on 25 May 2026. Article originally published in March 2026 — full rewrite.
Between 2023 and 2025, the framing of Perplexity as a direct competitor to Google structured part of the discourse on the evolution of online search. The frame was convenient: a conversational, citational, transparent challenger, set against a historic engine cluttered with adverts and sponsored blocks. This grid produced useful analyses, but it also misled on the nature of the shift under way.
After Google I/O 2026 and the official Google Search Central doctrine of 15 May 2026[1], this comparative framework deserves to be replaced by a more accurate reading: Perplexity and Google are not in a duel, they correspond to two complementary search regimes that B2B users activate depending on the moment and the nature of their need. This note sets out what this reading changes for a company that wants to understand its visibility.
Two regimes, not a duel
A professional user does not use a single search engine across all their needs. They switch from one regime to another according to what they are looking for.
When they look for a site whose existence they know, contact details, opening hours, a one-off factual answer, they use Google the same way they did ten years ago. The regime is navigational or transactional; no conversational engine replaces that need.
When they prepare a more complex decision — qualifying a market, comparing solutions, checking an actor's position on a subject, synthesising several sources — they increasingly switch to a conversational environment: Perplexity, ChatGPT in search mode, Claude with web access, or the now widely deployed Google AI Mode version. The regime is deliberative or comparative; the synthesised response has a usage value the list of links does not provide.
This distinction is not a methodological novelty of the firm; it is observable in measured usage and has been described by several professional publications since 2024. What is new after May 2026 is that Google has explicitly integrated the conversational regime into its own interface, with the extended AI Mode and AI Overviews now activated on a growing share of queries. The boundary between Google and competing conversational engines has therefore become porous in the user experience.
What each environment reveals about a company
For a company measuring its visibility, each environment reveals something different.
Google in classical mode reveals how a company positions itself in the organic competition for precise queries. It is a stable, long-measurable piece of data, and it remains relevant. A poor organic position on a strategic query remains a problem in 2026 as in 2016.
Google AI Mode and AI Overviews reveal how Google's generative components select and synthesise sources on a query. These components rely on the underlying index — what is in the index can be cited, what is not cannot be — but they operate a tighter selection than the ranking of the first ten results. A company can be in position six on the organic list and not be cited in the AI Overview, or be in position twelve and be cited. This new metric is now unavoidable.
Perplexity reveals a variation of the same property, in an environment that relies on its own source-selection strategy, with a citational transparency that remains its distinguishing marker. Perplexity citations include clickable links to the sources, which makes it a measurable acquisition channel, unlike generative models that mention a company without necessarily redirecting to its site.
ChatGPT and Claude, when they operate in search mode, reveal yet another variation: their source selection relies on their own systems, distinct from both Google and Perplexity. A company can be cited differently across the five major generative environments for the same query. It is this inter-environment gap that constitutes the basis of a structured measurement of citability.
What Google I/O 2026 did to the distinction
The main lesson of Google I/O 2026, alongside the doctrinal Search Central publication of 15 May, is that Google does not leave the conversational terrain to its competitors. AI Mode is extended and integrated more deeply into the search interface. The search agents introduced at I/O 2026 — capable of executing tasks on behalf of the user, from product comparison to appointment booking — represent a new layer of usage[2].
For a company, this evolution has a practical consequence. Thinking visibility by distinguishing "Google engine" and "competing conversational engines" is obsolete. The right distinction is now between search regimes — navigational, informational, deliberative, comparative, transactional — which execute partly in Google and partly in alternative environments, with growing overlaps.
Implications for the measurement of citability
Measuring a company's citability in 2026 therefore no longer reduces to observing Google. It supposes a protocol covering several response environments, on business-anchored queries, according to an explicit evaluation grid.
The firm has stabilised for this purpose a proprietary instrument, the Score GEO™, which runs on five generative environments (ChatGPT, Claude, Perplexity, Gemini, Mistral) according to a seven-theme grid. The full method appears in the Cahier MCVA n°1.
The point that matters for this note is not the MCVA method in particular, it is the necessity of a multi-environment protocol. A measurement conducted on Google only systematically under-estimates the diversity of situations in which a company is judged by the models. A measurement conducted only on a competing generative environment overstates the specificity of that environment and underplays the still considerable weight Google holds in effective practices.
The right strategic lens
For a Swiss B2B company wondering how to position itself after May 2026, the right lens is not to choose between a "Google SEO" strategy and a "presence on AI engines" strategy. This opposition no longer has any official doctrine to ground it. The right lens consists of recognising that users activate several search regimes according to their need, that each regime requires the same qualitative fundamentals (clarity, structure, sources, third-party authority, updates), but that they reveal partially different visibility situations.
The operational consequence is sober: continue to hold the SEO fundamentals with rigour, add a measurement of citability in generative environments on the queries that matter, and correct the gaps over time. The rest belongs to marketing, not to strategy.
Sources
[1] Google Search Central, Optimizing your website for generative AI features on Google Search, published 15 May 2026. developers.google.com/search/docs/fundamentals/ai-optimization-guide [↩]
[2] Google Blog, What's new in Search at Google I/O 2026, published 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.