McKinsey just published one of the clearest breakdowns yet of what's actually happening to advertising economics as AI takes over discovery and purchase decisions.
AI is reshaping advertising economics from discovery through transaction. McKinsey lays out four possible trajectories — and the common thread is that execution alone won't hold value anymore.
McKinsey just published one of the clearest breakdowns yet of what's actually happening to advertising economics as AI takes over discovery and purchase decisions. A few things stood out to me:
→ Search is being consolidated into AI environments. Over 50% of Google searches now show an AI overview, and up to half of open-web search traffic could be at risk by 2030.
→ AI is already mediating the transaction, not just the discovery. Shopping agents are ranking, comparing and in some cases completing purchases — McKinsey estimates 10-35% of e-commerce transactions could soon be initiated or influenced by AI-native experiences.
→ Value is concentrating in platforms that bundle data, measurement and transactions in one place. That's squeezing the traditional intermediaries (agencies, ad networks, exchanges) who historically connected buyers and signals separately.
→ 42% of advertisers already flag reliance on 'black box optimisation' as a top risk. You lose visibility into where your money actually lands.
The most useful part of the report isn't the trend-spotting though — it's the four possible trajectories it lays out for where this heads:
→ AI evolution: usage keeps growing but humans still make the final call and most budget stays in existing channels
→ Closed-web shift: discovery consolidates inside AI answers and brands compete for visibility inside the response itself
→ Autonomous delegation: agents research, recommend and transact on a person's behalf; influence shifts from what a consumer clicks to the model doing the choosing on their behalf
→ Curated ecosystem: premium publishers and retailers pool authenticated audiences behind shared measurement, as a hedge against consumer scepticism of AI-generated answers.
McKinsey is careful not to call a winner and poses that different categories will likely land in different places on this spectrum at different speeds. But the direction of travel across all four is the same: execution alone won't hold value anymore and the players who win are the ones building something hard to copy.
For brands, the practical translation is simple: being surfaced, recommended and selected by AI systems is now a distinct discipline from traditional paid media. If your product data, structured feeds and citations aren't built for machine evaluation, you're invisible to a growing share of the funnel — no matter how good your creative is.
That's the reason I built MARK: to show brands how they show up in AI models, who gets chosen over them and why.
- —AI is consolidating search, mediating transactions and concentrating value in bundled platforms — traditional ad intermediaries are being squeezed.
- —McKinsey's four trajectories all point the same way: execution alone won't hold value; the winners build something hard to copy.
- —Being surfaced, recommended and selected by AI is now a separate discipline from paid media — product data, feeds and citations must be machine-readable.