What's actually changing

For two decades, Google was the front door for nearly every commercial intent on the web. Type a question, get ten blue links, click one. The economics of the open web were largely shaped by that flow.

That flow is now bifurcating. A meaningful share of category-research queries are getting answered by AI assistants — ChatGPT, Claude, Perplexity, Google's own AI Overviews — without the user ever clicking through to a website. Other queries still flow through classic search.

Both the "Google is dying" and "AI search is hype" framings are wrong. What's happening is more interesting: the discovery surface is fragmenting, and businesses that adapt to multiple surfaces simultaneously are pulling ahead of businesses that don't.

Marketing MIX watches both surfaces for the clients we run. AI-search visibility and classic Google rankings move on different clocks: the same client can be early on one while the other stays flat, and most qualified traffic can still arrive through classic Google while it happens. Both are true at once.

Where each surface wins

Roughly:

| Query type | Primary surface in 2026 | |---|---| | "Best [category] for [specific use]" — comparison research | AI assistants (ChatGPT, Claude, Perplexity) | | "How do I [specific action]" — how-to | Google + AI Overviews + YouTube | | "[Brand name]" — navigational | Google, unchanged | | "[Product] price" — transactional | Google, unchanged | | "[Service] near me" — local | Google Maps, unchanged | | "Explain [concept]" — educational | AI assistants, increasingly | | "[Industry topic] news" — current events | Google + X/Twitter + Perplexity for citations | | "What is [obscure technical term]" — definitional | AI assistants, increasingly |

The pattern: research and education shift toward AI; transaction and navigation stay with Google; local stays with Google Maps; the middle (consideration, comparison) is the most contested.

What this means for marketing investment

The marketing function in most businesses is still organized around classic SEO and paid search as the discovery channels. That allocation made sense in 2020. In 2026 it's underweighting the new surfaces.

A reasonable reallocation for most businesses:

  • 70% of search-discovery budget to classic SEO + paid search (down from 95%).
  • 20% to AI-search visibility work (up from 0% in most businesses).
  • 10% to monitoring and adapting to surface shifts.

The exact ratio depends on category. B2B tech and professional services should weight harder toward AI-search visibility. Local services and high-volume e-commerce should weight harder toward classic Google.

What "optimizing for AI search" actually means

It's not a separate, mysterious discipline. It builds on top of classic SEO, with new emphasis on:

1. Entity binding

Every page that names your business should bind it to a clear category and location: "Marketing MIX, an international marketing studio with Ukrainian roots, headquartered in Ottawa and working across Canada, Ukraine, Germany, and France..." This phrasing is what LLMs extract when building their internal knowledge graph.

2. Structured factual claims

Pages with specific numbers (prices, durations, headcount, year, count) get cited more often than pages with vague descriptions. The 2024 trend of "AI prefers structured data" continues to play out.

3. FAQ schema at the bottom of every page

FAQPage schema is one of the highest-leverage technical SEO investments for AI citation. It's also the cheapest. Most sites we audit don't have it.

4. Author bylines with credentials

Anonymous "Our Team" content gets cited less than content with named, credentialed authors. Editorial bylines aren't just SEO theatre anymore; they're a citation signal.

5. Citation-friendly content structure

H1 stating one claim, H2s containing evidence, definitions in DefinedTerm schema, FAQs in FAQPage. This structure is more extractable than long-form prose.

6. Cross-platform consistency

Your brand entity needs to be consistently described across LinkedIn, Crunchbase, Wikidata (where applicable), industry directories, and your own site. LLMs cross-reference. Inconsistency creates ambiguity that lowers citation confidence.

We cover this in depth at /promotion-in-ai-search-and-llm-systems.

What the platform shifts mean for SEO investment timing

The economic logic for SEO has actually improved despite the disruption, with one caveat:

  • Top positions matter more. AI assistants citing one source weight rankings more heavily than a ten-result page did. Being #1 is worth more relative to being #5 than it used to be.
  • Topical authority matters more. AI assistants cite sources they've seen consistently across many related queries. A site with deep, consistent coverage of a topic outperforms a site with a single ranking page.
  • Brand-search and direct-traffic matter more. AI assistants reinforce strong brands and ignore weak ones. Brand-building investment pays back in citation behavior, not just direct conversions.

The caveat: bottom-half-of-page-one rankings matter less than they used to. A ranking that was producing a small amount of traffic in 2020 may now produce almost none, because the AI summary at the top captures the click.

What we actually do for clients

A typical engagement in 2026 includes:

  1. Classic SEO audit and remediation — site speed, schema, content quality, technical foundation.
  2. AI-search visibility audit — what does ChatGPT, Claude, Perplexity say about your brand, your category, your competitors? What's the current citation gap?
  3. Content restructure — top pages rewritten for both classic and AI-search citation.
  4. Entity-binding work — across owned and earned surfaces.
  5. AI mention monitoring — weekly automated probes of major models for your brand and category queries.
  6. Ongoing iteration — quarterly review of what's working and what isn't.

The cost of this work, for a mid-market business, is typically €4,800–€18,000 for the audit-and-restructure phase, then €2,800–€8,500/month for ongoing work.

What not to do

A few traps we see clients fall into:

  • Hiring a "GEO specialist" agency that ignores classic SEO. The two are not separable. An agency that doesn't do both well isn't ready.
  • Building content explicitly for AI ingestion. Content needs to serve humans first; LLMs extract from human-readable content best. AI-content-first sites end up with prose that humans dislike and LLMs increasingly down-weight.
  • Chasing every new "AI search engine." Optimize for the dominant ones (ChatGPT, Claude, Perplexity, AI Overviews). The next ones to emerge will mostly use overlapping signals.
  • Ignoring AI search because "we already rank well on Google." The Google rankings won't necessarily protect you if your category-research queries shift to AI assistants.

Related

For our AI-search visibility practice: /promotion-in-ai-search-and-llm-systems. For classic SEO foundation: /seo-promotion. For the broader retrospective on AI's industry impact: /blog/it-business-sharks-demand-moratorium-on-development-of-artificial-intelligence.

Written by Maksym Stepanenko, founder of Marketing MIX. Last reviewed: 2026-05-13.