Retail media has been the darling of CPG marketing budgets for the last three years. But here’s the uncomfortable truth: the entire model is built on a consumer behavior that AI is actively eroding — search-and-scroll discovery.
If AI shopping agents start completing purchases without a human ever seeing your sponsored placement, your retail media investment doesn’t just underperform. It disappears.
That’s not a reason to panic. It’s a reason to plan — right now, before the budget cycle locks you in.
What Agentic Commerce Actually Means for Retail Media
Agentic commerce is when AI acts on behalf of a consumer — browsing, comparing, and buying without the human doing any of it manually. Think of a user telling their AI assistant to “reorder my usual pantry items” and the agent completing the transaction based on past behavior, pricing signals, and brand data — with zero human browsing involved.
The implication for retail media is severe. Sponsored product placements, banner ads, and even promoted search listings inside retail networks only work when a human is looking at a screen. Remove that moment of visual discovery, and you’ve removed the entire premise of how retail media generates return.
This isn’t a future problem. Perplexity has already launched a buy-now feature inside its AI search interface. The behavior is being built right now.
Why CPG Brands Are Most Exposed
CPG brands have poured enormous budgets into retail media networks — Walmart Connect, Amazon DSP, Kroger Precision Marketing — because the targeting looked precise and the proximity to purchase looked unbeatable.
And it was, under the old model.
The problem is that most CPG retail media programs are optimized almost entirely around shelf-adjacent visibility. That works when a consumer is actively browsing. It does almost nothing when an AI agent is executing a replenishment order based on a preference profile it already built.
Here’s what makes CPG specifically vulnerable:
- High repurchase rates mean agentic systems will default to habit, not discovery
- Low brand switching intent means consumers won’t override the AI’s default choice
- Commoditized categories mean price and availability signals — not creative — will drive agent decisions
- Thin margin structures mean there’s no room to outbid every AI-mediated placement that emerges
The brands that win in agentic commerce won’t be the ones who spend the most on placements. They’ll be the ones the AI already trusts.
What Performance Marketers Need to Do Right Now
This isn’t about abandoning retail media. It’s about not being 100% dependent on a model that a single behavioral shift can break.
Our take at Junction 37: the brands that come out ahead will be building on three fronts simultaneously.
1. Strengthen owned first-party signals.
If an AI agent is making decisions based on a consumer’s history, your goal is to be in that history. Direct purchase data, loyalty program depth, and email engagement are all inputs that shape an AI’s understanding of what a customer “usually buys.” [Our performance media work](/services/performance-media) increasingly treats first-party data infrastructure as a media channel in its own right.
2. Optimize for AI-readable brand attributes.
Structured product data — ingredients, certifications, sustainability claims, reviews — is becoming the new ad unit. AI agents pull from this data to make recommendations. If your product data is incomplete, inconsistent, or buried, you’re invisible to the agent before the consumer even enters the picture.
3. Diversify discovery budgets before you’re forced to.
Retail media still works today. But allocating 70%+ of a media budget to a single network’s sponsored placements is a concentration risk that most brands haven’t stress-tested. Our strategy team is actively helping CPG brands model what a post-placement discovery mix looks like — and it looks very different from the 2023 playbook.
The Agency Responsibility Nobody Is Talking About
Most holding company agencies will keep optimizing retail media ROAS because that’s what the reporting dashboard rewards. They won’t tell you the model has structural risk because that conversation is uncomfortable and the fees are good.
Independent performance agencies don’t have that incentive. Our job is to tell you what’s coming and build the plan before the disruption hits your Q3 numbers.
Retail media isn’t dead. But blind faith in sponsored placements — without a plan for what replaces them — is a liability.
FAQ: Retail Media and Agentic Commerce
What is agentic commerce?
Agentic commerce refers to AI systems that complete shopping tasks autonomously on behalf of a user — including browsing, comparing, and purchasing — without requiring the consumer to manually interact with a retail interface.
Will agentic commerce kill retail media networks?
Not immediately, but it will significantly reduce the value of placement-based retail media over time. Networks that adapt by offering AI-agent-accessible product data and preference-based targeting will survive. Those that don’t will see declining returns as consumer browsing behavior shifts.
What should CPG brands do to prepare for this shift?
CPG brands should invest in first-party data infrastructure, ensure their product data is structured and AI-readable, and diversify media budgets beyond sponsored placements in retail networks.
How soon will this affect my retail media performance?
Early signals are visible now — AI shopping interfaces with purchase capability already exist. The behavioral shift will likely accelerate in 2025–2026 as AI assistants become more embedded in daily consumer routines. Planning should start before this quarter’s budget is locked.
Ready to build a retail media strategy that holds up past the next platform shift?
Talk to Junction 37’s performance media team — we work exclusively with CPG, DTC, and purpose-driven brands who need a plan, not a pitch deck.
Chris Pyne, Founder, Junction 37 – 30+ Years in Performance Media