Your customers haven't left Google for Perplexity: they split their questions across several search regimes. Simple queries — a known website, an opening time, a contact detail — stay on classic Google; substantive questions — comparing providers, sizing up a market, preparing a decision — increasingly shift toward conversational environments, including Google's own AI Mode. The question that actually matters for your business is no longer "which engine is winning?" but "am I present in every regime where my customers look for me?"
Article published in March 2026, fully rewritten on May 25, 2026 following Google I/O and the Search Central doctrinal publication.
A duel that isn't one
Between 2023 and 2025, framing Perplexity as Google's direct competitor shaped much of the discourse on online search. The framework was convenient: a conversational, citation-driven, transparent challenger against a legacy engine cluttered with ad blocks. This lens produced useful analysis. It also misled people about the true nature of the shift underway.
A professional user doesn't rely on a single engine for all their needs. When looking for something they know exists, they use Google exactly as they did ten years ago; no conversational tool replaces that reflex. When preparing a more complex decision, they switch to an environment where the answer is synthesized: Perplexity, ChatGPT in research mode, Claude with web access, or Google's now widely deployed AI Mode. The synthesis then carries a usage value that a list of links doesn't provide.
This distinction between regimes isn't something I invented; it's observable in usage patterns and has been described by several trade publications since 2024. What's new after May 2026 is that Google has folded the conversational regime into its own interface, with an expanded AI Mode and AI Overviews now active on a growing share of queries. The boundary between Google and its conversational competitors has become porous in your customers' actual experience.
What each environment reveals about your business
Each search terrain illuminates a different facet of your visibility, and none tells the whole story.
Classic-mode Google reveals your position in organic competition on precise queries. It's a stable, long-measurable, and still relevant metric: a poor position on a strategic query remains a problem in 2026 just as it was in 2016.
Google AI Mode and AI Overviews reveal how Google's generative components select and synthesize sources. They draw on the index — what isn't there can't be cited — but apply a tighter selection than the top ten results. A business can rank sixth in the organic list and be absent from the AI Overview, or rank twelfth and get cited. That gap deserves to be tracked just as closely as ranking itself.
Perplexity reveals a variation on the same property, with a distinctive marker: its citations include clickable links to sources. That makes it a measurable acquisition channel in your analytics, unlike models that mention a business without linking back to its site. ChatGPT and Claude, in research mode, rely on their own selection systems, distinct from both Google and Perplexity. A single query can therefore produce five different representations of your business — and it is precisely this gap between environments that underpins a structured measurement of citability, the one I describe in AI visibility: does your SME show up in ChatGPT?.
What Google I/O 2026 changed
The main takeaway from Google I/O 2026, held four days after the May 15 Search Central doctrinal publication[1], is that Google isn't ceding conversational territory to its competitors. AI Mode has been expanded and integrated more deeply into the search interface; the research agents unveiled at the event — capable of performing tasks on the user's behalf, from product comparison to booking appointments — add yet another layer of usage[2].
The practical consequence for an executive: thinking about visibility as "Google's engine" versus "AI engines" is now obsolete. In 2026, the right unit of analysis is no longer the engine but the search regime — navigational, informational, deliberative, comparative, transactional — each running partly within Google and partly within conversational environments, with growing overlap between the two.
How do you measure without picking the wrong terrain?
Measuring your visibility in 2026 can no longer be reduced to watching Google. A measurement conducted solely on Google underestimates the range of situations in which models judge your business. A measurement conducted solely on Perplexity or ChatGPT overstates that environment's specificity and understates Google's still-considerable weight in real-world practice. What's needed is a multi-environment protocol, run on business-anchored queries, following an explicit framework.
I've stabilized an instrument for this purpose, the Score GEO™, which runs across five generative environments — ChatGPT, Claude, Perplexity, Gemini, Mistral — using a seven-theme framework. The full methodology is set out in MCVA Cahier No. 1. But the point that matters here goes beyond my own instrument: whatever protocol you choose, it must cover several environments, or else it describes one patch of ground while claiming to describe the whole landscape.
The right strategic lens follows from this. Every regime demands the same fundamentals — clarity, structure, sources, third-party authority, freshness — and those fundamentals are built first on your own website: that's the work of a site designed to be cited, which I detail elsewhere and which serves every regime at once. From there, measure citability on the queries that matter, and correct gaps over time. I've placed this trade-off within an SME's overall roadmap — visibility, tools, compliance — in What should a Swiss SME do about AI in 2026?. Everything else is marketing, not strategy.
Key takeaways
— Your customers alternate between search regimes: classic Google for the simple stuff, conversational environments for the complex — including Google's own AI Mode. — Each environment reveals a different facet of your visibility; none is sufficient on its own. — The measurement that matters is multi-environment: the same fundamentals apply everywhere, but citability situations differ partly and need to be tracked over time.
FAQ
Will Perplexity replace Google?
Nothing suggests it will. Observed usage shows a split by regime: Google keeps navigational and transactional searches, while conversational environments capture a growing share of deliberative and comparative searches. Google has, in fact, folded this regime into its own interface with AI Mode.
Do I need to optimize differently for Perplexity versus Google?
No. Google's May 15, 2026 doctrine and observable practice converge on this: the same fundamentals — useful content, clean structure, transparent sources, built-up authority — serve every environment. What differs is the selection each system applies, and that's something you measure rather than optimize separately for.
Do citations in Perplexity drive traffic?
Perplexity displays clickable links to its sources, making it an acquisition channel you can observe in your analytics. Other environments sometimes mention a business without a link: the value there is reputational — being recommended — rather than directly measurable in visits.
Ranking well on Google but absent from AI answers: what should you do?
First, establish the gap objectively through measurement across several environments and queries, not an isolated test. Then fix what explains the absence — most often a shortfall of citable content on your website, of sources, or of third-party authority — and re-measure on a steady cadence to track the trajectory.
Where do you show up, and where don't you? The Score GEO™ pre-audit measures, free of charge, where your business appears — and doesn't appear — in the answers from ChatGPT, Claude, Perplexity, Gemini, and Mistral. Quantified results within a few days. Request a free pre-audit
Sources
[1] Google Search Central, Optimizing your website for generative AI features on Google Search, published May 15, 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 and founder of MCVA Consulting SA, an augmented agency based in Valais, Switzerland. Fifteen years serving major international brands, now working directly with Swiss SMEs. Background.