There is a specific kind of forecast that sounds wildly speculative until the moment it stops being speculative. The agentic commerce market growing from $5 billion in 2025 to $145 billion by 2029 sounds like one of those forecasts.
It isn’t.
The data from the past 30 days makes clear that the transition is not waiting for some future technological breakthrough. It is happening now, in product launches, infrastructure investments, and consumer behavior shifts that are already in motion. Meta just launched a Business Agent designed to keep entire customer conversations inside its ecosystem. Google is building a Universal Commerce Protocol for cross-platform AI shopping. Pinterest committed $4 billion to AWS to power visual AI search. And 50% of consumers are already using AI for the research phase of their purchasing journey.
The marketing playbook that worked even six months ago is being quietly retired. Here are the seven developments that prove it.
1. Meta’s New Business Agent Could Disintermediate Every Merchant Site It Touches.
Meta has launched a customer-facing AI Business Agent designed to handle the entire customer journey inside its messaging ecosystem: product recommendations, lead qualification, appointment booking, and increasingly, transaction completion. The strategic implication is significant and uncomfortable for most brands.
For two decades, social media platforms operated on a “click-out” model. The platform served the discovery and engagement layer, then sent traffic to the brand’s own website where the conversion happened. The Business Agent flips that entirely. Meta is now positioning itself as a “stay-in” conversion engine, where the consumer never leaves the Meta environment to complete the purchase.
This sounds like a convenience feature. It is actually a competitive shift. Every conversation that completes inside Meta is data Meta owns, a relationship Meta controls, and a customer experience the brand can no longer fully shape. Brands that have spent years optimizing their post-click experience now face a choice: integrate deeply with Meta’s agentic infrastructure and lose some direct relationship control, or invest in alternative channels and risk losing reach within the world’s largest social ecosystem.
2. Google Is Building an “HTTP for AI Shopping.” Standards Are About to Matter More Than Tactics.
Google’s development of a Universal Commerce Protocol represents one of the most consequential infrastructure developments in digital commerce in years. The concept is straightforward: an HTTP-like standard that allows AI agents to negotiate, compare, and purchase across platforms with consistent data attributes.
The marketing implication runs deep. In an HTTP-driven web, brands optimize for how their pages appear to humans. In a protocol-driven agentic web, brands need to optimize for how their products are described to AI agents. Product data attributes, structured pricing, real-time inventory signals, comparison-friendly specifications, all of these become the new “creative” that determines whether an AI agent recommends your product or your competitor’s.
“Brands must adopt standardized data attributes to ensure AI agents can negotiate and purchase on behalf of users.”
The brands investing now in clean, structured, machine-readable product data are building an infrastructure advantage that will compound as agentic commerce scales. The brands still treating product descriptions as marketing copy rather than agent-readable signals are about to discover that the AI agents are not reading them carefully.
3. Pinterest Just Bet $4 Billion on Visual AI Search. This Is the Quietest Big Deal of 2026.
Pinterest has committed $4 billion in AWS infrastructure spend through 2031, specifically targeting its Performance+ ad suite and Bedrock-powered visual search capabilities. In a year dominated by AI infrastructure announcements, this one has received relatively little attention. It deserves significantly more.
Pinterest already processes more than 80 billion searches per month with approximately 50% commercial intent, a ratio that exceeds virtually every other platform in the ecosystem. A $4 billion infrastructure commitment dedicated to enhancing visual search and AI-driven recommendations is positioning Pinterest as a high-intent commerce discovery platform at a scale most marketers have not yet absorbed.
For CPG, retail, home, beauty, and fashion brands, the implication is direct. Pinterest is building the visual equivalent of an AI answer engine, where users describe what they want and the platform’s AI surfaces specific products. Brands that have been treating Pinterest as a brand awareness channel are missing what the platform is actually becoming: a structured, high-conversion alternative to traditional search for visual product categories.
4. The Search Spend Numbers Tell a Bleak Story for Google. And a Bleaker One for YouTube.
Here is what the agency expert data actually shows. Google Search ad spend grew 8% to 12% year-over-year in Q2 2026. The growth was driven primarily by cost-per-click inflation, not by volume expansion. In other words, advertisers are paying more for the same number of clicks, not benefiting from more search activity.
The picture for YouTube is more troubling. Some agency portfolios are showing a 5% year-over-year decline in YouTube spending, as budgets shift to TikTok, Twitch, and Roku. The largest video advertising platform in the world is losing share to a more fragmented ecosystem of attention.
The underlying cause is the same in both cases. AI Overviews and conversational AI platforms like ChatGPT and Perplexity are capturing research-style queries that previously fed traditional search and video discovery. Consumers are increasingly getting their answers without ever clicking through to a search results page or being served a pre-roll ad.
For marketers, the question is no longer whether to diversify away from Google’s ecosystem. It is how quickly the diversification can happen. The brands that have spent the last decade optimizing primarily for Google Search now face the same kind of platform dependence risk that publishers faced with Facebook five years ago.
5. Counterfeit Brand Ads Are Becoming a Platform-Level Crisis.
A trend surfaced in this period’s research that deserves more attention than it has received. Sophisticated counterfeit operations are using targeted ads on TikTok and Instagram to direct users to “clone websites” that mimic legitimate brands. The fraudsters are using the same targeting algorithms that authentic marketers depend on, exploiting them to reach high-intent consumers who think they are buying from a brand they trust.
The consequences are escalating quickly. Consumer council lawsuits against platforms for failing to screen fraudulent content are increasing. Legitimate brands are losing revenue, customer trust, and brand equity to clone operations they cannot easily shut down. And the targeting precision that makes social media advertising effective is the same precision that makes social media counterfeit operations devastating.
For brand-side marketers, the implication is operational. Brand protection investment, in monitoring tools, takedown infrastructure, legal response capacity, is no longer a security function. It is a marketing function. The brands that treat counterfeit defense as a routine cost of doing business on social platforms are the ones losing customers to fraudsters. The brands building genuine vigilance into their marketing operations are protecting both their revenue and the consumer trust that makes their marketing work in the first place.
6. The “Agent of Agents” Era Just Began. Agencies Will Look Completely Different in Two Years.
The launch of HAWK, an AI agent purpose-built for social media advertising operations, represents a small but telling signal of where digital marketing operations are heading. HAWK handles reporting, optimization decisions, asset iteration, and other repetitive ad operations tasks that historically required human agency labor.
This is the start of what some analysts are calling “the industrialization of the ad supply chain.” The next phase of digital marketing is not just AI helping marketers do their existing jobs better. It is AI agents handling the work of an entire agency function, with humans shifting upward to strategy, creative direction, and orchestration of multiple specialized AI agents.
The implication for agency-brand relationships is significant. The economic model that priced agency services based on hours of human execution is being disrupted by AI agents that can perform many of those hours for a fraction of the cost. The agencies that survive this transition will be the ones that reframe their value around strategic judgment, creative excellence, and AI orchestration, not around the volume of human labor they can apply to a problem.
7. The First-Party Data Imperative Just Became Existential.
The agentic commerce era has elevated first-party data ownership from “important” to “existential.” When AI agents are mediating discovery, comparison, and increasingly transaction itself, the brands that own deep relationships with their customers and have clean data on those relationships are the ones that retain pricing power and customer lifetime value. The brands that depend on platform intermediaries to identify, target, and convert their customers are renting access at terms the platforms increasingly control.
Walmart and Amazon are gaining share specifically because they offer advertisers “closed-loop” measurement, the ability to see direct impact of ad spend on actual sales. The data fabric they have built is the asset, and the advertising performance is the output. Brands without comparable first-party data infrastructure are competing on a different playing field.
“As search shifts to agents, the most successful brands will be those that own their first-party data ecosystems.”
The strategic question for every CMO in 2026 is uncomfortably direct. What is your brand’s plan for owning the customer relationship when the discovery layer, the comparison layer, and the transaction layer are all being mediated by AI agents you don’t control? Brands that have an answer are investing accordingly. Brands that don’t are buying themselves a more expensive future where every customer interaction is brokered by a platform with its own incentives.
The Pattern Across All of It
Every development in this period’s data points toward the same underlying shift. The locus of value in commerce is moving from the customer’s screen to the AI agent that operates on the customer’s behalf. The marketing function that succeeded in optimizing for human attention now has to learn to optimize for agent recommendation, structured data clarity, and platform integration depth.
This is not a technology change layered on top of existing marketing practice. It is a re-foundation of what marketing actually does. The skills that mattered most for the past decade, audience targeting, creative iteration, conversion rate optimization, are not becoming obsolete, but they are becoming insufficient. The new skills, structured data architecture, agent visibility optimization, first-party data ecosystem design, AI orchestration, are not yet widely held in most marketing organizations.
The brands that recognize this gap and start closing it now will have a meaningful competitive head start by 2027. The brands that wait until agentic commerce is undeniably the dominant model will be trying to build infrastructure while their competitors are already deploying it.
– Manpreet Jassal
Leave a Reply