What is agentic media buying and does it actually work for Shopify brands?

Agentic media buying means an AI system makes campaign decisions — bids, placements, budget shifts — autonomously, without a human approving every move. Tools like Meta Advantage+ and Google Performance Max are real, working versions of this. They've driven genuine results for brands we manage. The problem is that a lot of platforms are now calling themselves 'agentic' when they're really just rules-based automation with a GPT wrapper on top. The way to tell the difference: ask for a decision log. If the platform can't show you what it changed, when, and why — and what happened to your ROAS as a result — it's not truly agentic. Don't pay an agentic premium for a black box.

How is the AI Overview trend on Google actually affecting Shopify store traffic?

It's affecting informational queries the most. If you've built content around 'how to' topics, ingredient explainers, or category education, you're likely seeing impressions hold steady but clicks drop. AI Overviews answer the question and keep the user on Google. What's still sending clicks: high-intent transactional queries, brand-specific searches, and product comparison content that AI can't fully satisfy. Our recommendation is to audit your Search Console data — filter by query type and look at your click-through rate trend over 90 days. Then shift content investment toward queries where a human genuinely needs to click through to get the full answer.

Why does the open web funding crisis at Cannes affect my DTC ad costs?

As AI tools absorb more queries, publishers lose traffic and therefore lose ad inventory. Less inventory with the same advertiser demand means CPMs rise. If your awareness strategy relies on display advertising, sponsored content on media sites, or retargeting audiences built from third-party publisher pixels, you're buying into a shrinking, more expensive pool. It's not a cliff-edge collapse — it's a slow squeeze. The brands that feel it hardest are those who haven't built owned channels to compensate. Your email list, SMS list, and direct search traffic are your insulation against this.

What should I actually prioritize in my Klaviyo setup to protect against AI disruption?

Focus on depth, not breadth. Most brands have a welcome series and an abandoned cart flow — and that's it. The real revenue is in flows that activate based on behavior: browse abandonment, post-purchase cross-sell, win-back sequences, and VIP segmentation. We audit brands regularly where 60-70% of potential email revenue is sitting in unbuilt or underoptimized flows. AI disruption to paid and organic channels makes this gap more painful because you can't easily buy your way back to those customers. Own the relationship before you need it. A well-structured Klaviyo program should be driving 30-40% of total brand revenue.

Are AI shopping agents actually crawling my Shopify store right now?

Yes. Perplexity's shopping feature, ChatGPT's browsing capability, and browser-based AI extensions are all reading product pages, parsing structured data, and making recommendations. Apple Intelligence is building toward this as a native experience. This means your product data quality now directly affects whether an AI agent recommends you or a competitor — not just whether a human likes your page. Thin descriptions, inconsistent specs, and missing structured data all hurt you with AI agents. Audit your product pages for completeness: clear pricing, specific materials or ingredients, sizing, who the product is for, and honest reviews. That's what both humans and AI agents need.

How do I know if an 'AI-powered' ad platform is worth paying for?

Ask three questions before signing any contract. First: can you show me a decision log — what the system changed, when, and why? Second: how does your optimization logic handle new creative with no historical data? Third: what exclusions, brand safety controls, and audience overrides can I set manually? If the answers are vague, the platform is probably rules-based automation with AI branding on top. Genuinely useful AI ad tools make their decision-making transparent, improve with better inputs you provide, and give you clear override controls. If they can't answer those three questions specifically, you're paying a premium for a black box — and your ROAS will reflect that.

AI Money vs Agentic Hype: What DTC Brands Must Do Now

Every year Cannes Lions produces a lot of noise. This year it produced one genuinely useful signal.

Publishers showed up on the Croisette with two hard questions on their agenda: who actually pays for the open web in an AI world, and whether agentic media buying is a real shift or just a rebadged ad-tech tax. Those questions sound like publisher problems. They're not. They land directly on your Shopify store's customer acquisition budget — and if you're not paying attention, you'll feel the impact before you understand why.

I'm going to tell you what I think is actually happening, what we've seen across our client base, and what you need to do about it right now.

The "Open Web" Funding Crisis Is Your Ad Cost Problem

Here's the short version: AI is eating traffic.

Google's AI Overviews answer questions before users click. ChatGPT, Perplexity, and a dozen other tools surface product recommendations without sending anyone to a publisher's article or your product page. Publishers are seeing referral traffic crater. That means their ad inventory is shrinking. Less inventory, same advertiser demand — your CPMs go up.

We've been tracking this across clients for 18 months. The brands that leaned entirely on top-of-funnel content from third-party publishers for awareness — think sponsored editorial, display retargeting from news sites — have watched their reach-per-dollar deteriorate meaningfully. It's not catastrophic yet. But the trajectory is clear.

If you're still building your ecommerce customer acquisition strategy around the assumption that the open web is a stable, affordable discovery layer, that assumption is breaking.

Agentic Media Buying: Real Shift or Ad-Tech Tax?

I'll take a clear position here: agentic media buying is real, but most of what's being sold to DTC brands under that label right now is not.

Genuine agentic buying means an AI system makes bid, placement, and budget allocation decisions autonomously — learning in real time, adjusting without a human approving every move. Meta's Advantage+ and Google's Performance Max are early, imperfect versions of this. They're not hype. They've moved significant revenue for brands we work with.

The hype is the layer of "AI-powered" DSPs and "autonomous creative optimization" platforms that have wrapped a rules engine in a GPT-flavored UI and tripled their SaaS fee. I've sat through six pitches in the last quarter from platforms in this category. None of them were doing anything materially different from a well-configured manual campaign.

The tell is simple: if the platform can't show you a clear decision log — what the system changed, why, and what happened to ROAS — it's not agentic. It's a black box with a marketing rebrand.

For a proper breakdown of ad creative measurement in DTC paid media, you need transparency at the creative and placement level. Agentic systems that can't give you that are charging you for autonomy you can't verify.

What AI Is Actually Doing to Your Acquisition Funnel Right Now

Let me be specific about three things we're seeing in our client accounts.

1. AI Bots Are Touching Your Store Before Humans Do

This one surprises most brand founders. AI shopping agents — the kind built into Perplexity, built into browser extensions, and increasingly into Apple's ecosystem — are crawling your Shopify store, reading your product data, and making recommendations to users who never visit your site.

We wrote about how Apple Intelligence and Siri are reshaping commerce earlier this year. The implication is that your product titles, descriptions, and structured data now serve two audiences: human shoppers and AI agents. If your product data is thin, inconsistent, or keyword-stuffed in the old SEO style, AI agents will deprioritize you.

There's also a protection angle. AI bots hitting your Shopify cart and checkout can distort your analytics, inflate your add-to-cart rates, and burn ad budget on retargeting audiences that were never real buyers. We've fixed this for several clients and the analytics cleanup alone changed how they allocated their media spend.

2. AI Overviews Are Reshinking Your Organic Funnel

If you've been building content to drive organic acquisition, you need to look at your Google Search Console data for the last 90 days and check impression-to-click ratios on informational queries. For most of our clients, impressions have held up. Clicks have dropped. AI Overviews are answering the question and keeping the user on Google.

Our 18-month data on Google AI Overviews tells a consistent story: transactional queries (high purchase intent, product-specific) are still sending clicks. Informational queries (how to, what is, best X for Y) are being absorbed by AI Overviews at an accelerating rate.

The correct response isn't to stop publishing. It's to shift your content investment toward queries where AI Overviews can't fully satisfy intent — deep comparison content, brand-specific reviews, UGC-heavy pages, and SEO strategies built for AI agent indexing.

3. First-Party Data Is Now Your Primary Moat

This is not a new point. But Cannes 2025 made it undeniable: the brands that own their customer data are dramatically less exposed to everything the AI transition is doing to paid and organic channels.

Your email and SMS list is the one acquisition and retention channel that no algorithm change, no AI Overview, no agentic bidding system can take away from you.

Channel	AI Disruption Risk	Your Control
Google Organic (informational)	High	Low
Meta Paid (manual)	Medium	Medium
Meta Advantage+ / agentic	Medium	Low-Medium
Open web display	High	Low
Email (owned list)	Very Low	Very High
SMS (owned list)	Very Low	Very High

If you haven't built a serious ecommerce customer data strategy, the window to do it cheaply is closing. List acquisition costs are rising as more brands chase the same first-party moat.

The DTC Brand Response: What We're Actually Recommending

We work with Shopify brands across beauty, supplements, fashion, and home. Here's the actual playbook we're running for the second half of 2025.

Double Down on Owned Channels

Your email and SMS programs need to be sophisticated — not just a welcome series and an abandoned cart flow. If your Klaviyo setup is still running the same five flows it was two years ago, you're leaving a significant retention revenue gap open. We've seen brands recover 15-25% of revenue that was leaking through unoptimized post-purchase and browse abandonment sequences alone.

The ecommerce email revenue percentage benchmark for a well-run program is 30-40% of total revenue. Most DTC brands we audit are at 15-20%. That gap is entirely capturable without spending more on acquisition.

Pressure-Test Your Paid Media Setup for the Agentic Era

If you're running Meta, you need to understand what Advantage+ is doing with your budget — not just accept the black box. Understanding your ecommerce Meta ads setup at the campaign architecture level matters more now than it did when manual targeting gave you more control.

The same applies to Google. Performance Max can be powerful or it can be a budget incinerator. The difference is in how you structure your asset groups, exclusions, and audience signals. Agentic systems reward better inputs — they don't forgive lazy campaign structure.

Build Product Data for AI Agents, Not Just Search Engines

Your product pages need structured, specific, accurate data. Not because Google's crawler needs it — because AI shopping agents need it to recommend you over a competitor.

Think about your product page above the fold: price, key benefits, materials, dimensions, who it's for. That's what a human needs to decide. It's also exactly what an AI agent parses to make a recommendation. The two audiences want the same thing. Give it to them.

Don't Chase Agentic Platforms You Can't Audit

I'll say this plainly: don't pay a premium for "agentic" ad platforms that can't show you a clear decision log. If they can't tell you what the system changed yesterday, why it changed it, and what happened to your cost per acquisition as a result — they're not agentic. They're expensive.

Your DTC marketing stack should be built on tools where you understand what's happening. Autonomy without transparency is just budget leakage with better branding.

The Cannes Conversation That Matters for Your Brand

The publishers at Cannes were asking who pays for the open web in an AI world. The honest answer is: fewer people than before, at higher rates, with less measurable return.

That's not doom. It's a structural shift that rewards brands who've built owned audiences, who understand their ecommerce analytics well enough to spot where AI is distorting their data, and who aren't over-indexed on channels they don't control.

The brands we're most confident about heading into Q4 aren't the ones with the biggest ad budgets. They're the ones with clean first-party data, a real email and SMS program, product pages that serve both humans and AI agents, and the discipline to audit every "AI-powered" tool before paying for it.

Agentic media buying is coming. AI is already reshaping discovery. The question isn't whether to engage with it — it's whether you're engaging on your terms or just absorbing the cost.

If you want to audit your current stack for AI exposure — where you're vulnerable, where you're wasting spend, and what to prioritize in H2 — that's exactly the kind of work we do at GOSH. Start with a proper AI in ecommerce marketing review of your channels before you commit another dollar to platforms making promises they can't prove.