Our Take
The thesis is sound but the essay is strategy-tier diagnosis, not a roadmap—McKinsey names the shift without showing who wins or how fast.
Why it matters
If AI agents become the primary discovery layer for purchase decisions, traditional ad placements (search, social feeds, display networks) lose pricing power to whoever controls the agent's recommendations. Advertisers need to know which platforms are already shipping agentic selection and which are still selling attention.
Do this week
Marketing leads: audit your top 5 media partners this week to confirm which ones have published agentic recommendation APIs or agent-integrated storefronts—don't assume legacy search and social properties are ready.
McKinsey's thesis on the agentic ad shift
McKinsey published a strategic brief arguing that advertising's value center is migrating. Historically, platforms sold reach and attention: eyeballs on a page, impressions served, clicks counted. The thesis claims AI agents will change the unit of value from visibility to outcome. As consumers increasingly use AI to discover and filter products rather than browsing directly, the platforms that shape what agents recommend will command pricing power over those that merely serve impressions to humans.
The framing is clear: three decision points (what is seen, what is selected, what is purchased) will be reshaped by agentic intermediaries. Whoever controls the agent's training data, ranking logic, or integration points controls the buyer's path.
The catch: this is diagnosis without deployment data
McKinsey is describing a future state that has not yet arrived at scale. No major ad platform has yet shipped agentic recommendation as a primary monetization surface. Google, Meta, and Amazon still make the vast majority of ad revenue from human-facing surfaces: search results, news feeds, product pages. ChatGPT, Claude, and Gemini have integration partnerships with some retailers and publishers, but these remain marginal channels for ad spend, not the primary discovery mechanism.
The strategic insight is real: if agents do become the primary discovery layer, the incentive to advertise on them will follow. But the article does not establish that agents are already capturing material share of consumer purchase decisions, nor does it name which platforms are closest to shipping agentic ad models in production. That gap matters because timing affects which advertisers and platforms are positioned to act.
The second-order risk is that platforms currently dominant in human-facing ads (Google, Meta) will simply integrate agents into their existing surfaces rather than creating new agentic revenue streams. If agents are embedded in search results or Instagram, the value still accrues to the incumbent platform, not to some separate agentic layer. McKinsey's framing assumes agentic discovery is separable from current platforms; that assumption is unproven.
What to do now
Stop waiting for a unified agentic ad market. Instead, run a quarterly audit of the platforms where your target buyers use agents. Map which ones already let you bid on or integrate with agent recommendations (Shopify's agent tools, Google's AI Overviews, Amazon's Q, etc.). Prioritize the channels where agents are already shaping selection. Ignore press releases about "agentic advertising strategies" without published integrations or spend capabilities. The platforms proving agentic discovery at scale will announce the ad product second, not first.