AI Shopping Channel Showdown 2026: ChatGPT vs Google AI Mode vs Perplexity vs Copilot, by Real Conversion Data
The hype versus what the data actually says
Every e-commerce Slack I’m in has been buzzing about AI shopping rewriting retail. I bought into it for a while too, until I read the peer-reviewed study published in Marketing Science (also up on SSRN). It covers August 2024 through July 2025: 973 e-commerce sites, 20 billion USD in revenue, more than 50,000 LLM transactions measured against 164 million transactions from every other channel. The headline number is humbling. Organic traffic from large language models is still under 0.2% of all e-commerce sessions.
Sit with that for a second. Your company has probably spent more hours in meetings about AI shopping than the channel is currently worth in actual revenue. The volume and the noise are wildly out of proportion, and that’s the baseline you have to keep in mind before reading any comparison.
The 0.2% is also lopsided inside itself. ChatGPT alone takes over 90% of LLM e-commerce traffic. The rest split thin: Perplexity around 4.1%, Gemini around 2.6%, Copilot around 2.1%. So “the four big channels” sounds like a four-way fight, but it’s really one giant and three slivers. Carry that ratio in your head when you allocate effort, because treating the four as equals means spreading yourself across rounding errors.
None of this says the channel is worthless. It grows fast and the intent is high, which I’ll get to. But anyone telling you to go all-in on AI shopping right now almost certainly hasn’t looked at the volume.
Four channels, side by side
Lined up in one table, the differences jump out:
| Dimension | ChatGPT | Google AI Mode | Perplexity | Microsoft Copilot |
|---|---|---|---|---|
| Share of LLM e-commerce traffic | over 90% | ~2.6% (Gemini basis) | ~4.1% | ~2.1% |
| Reach | ~900M weekly users | 1B+ monthly users | free for all US users | baked into Windows / Edge |
| Cost to participate | ad floor removed (from 2026-05-05); plus organic Shopping Research | gated by Merchant Center feed quality | free to join, ~5-min signup, zero fees | Shopify merchants auto-enrolled (opt-out) |
| Checkout model | feed ads + organic discovery, Conversions API + pixel | UCP-powered, US first then CA/AU/UK | Buy with Pro, Perplexity funds shipping | Copilot Checkout (live 2026-01-08) |
| Best fit | sellers chasing volume with feeds and content in place | sellers who already nailed Google Shopping feeds | small sellers wanting a free foothold | Shopify sellers, retailers with loyalty programs |
ChatGPT is the only one with both real volume and full infrastructure. As of May 5, 2026, its product-feed ads opened to all US advertisers with no minimum spend. The previous floor of 50K-250K USD is simply gone. Conversions API and pixel-based attribution launched alongside the ads, so for the first time you can actually measure what AI ad spend returns instead of guessing. There’s also an organic Shopping Research discovery surface. All of it sits on roughly 900M weekly users.
Google AI Mode has the biggest reach at 1B+ monthly users, with checkout running on the UCP protocol, rolling out in the US first and then Canada, Australia, and the UK. The catch is a hard dependency: it feeds off Merchant Center feed quality. A sloppy Google Shopping feed means you get almost nothing from AI Mode. For sellers who already maintain a clean feed, this is a free ride. For everyone else, it’s prerequisite homework.
Perplexity took the cheap-foothold route. It made shopping free for all US users, and its Buy with Pro merchant program is free to join with a roughly five-minute signup, zero merchant fees, and free shipping that Perplexity funds out of its own pocket. The traffic is only around 4.1%, but joining costs you almost nothing.
Microsoft Copilot launched Copilot Checkout on January 8, 2026. Shopify merchants were auto-enrolled with an opt-out, and Target became the first major retailer to plug in with loyalty integration. Traffic sits near 2.1%, but if you’re already on Shopify it’s essentially free exposure.
The conversion-metric trap: compared to what?
This is where people get talked into a narrative by a single number, so let’s be precise.
In the study, the conversion picture looked rough for AI. Affiliate links convert roughly 86% better than ChatGPT referrals, and organic search converts about 13% better than ChatGPT referrals. Read those two figures alone and the takeaway seems obvious: AI shopping converts poorly, skip it.
The same study tells the other half, though. ChatGPT conversion is climbing month over month, and revenue per session is rising too. That’s a curve still on its way up, not a mature plateau. Judging a growing channel by today’s absolute numbers against established channels stacks the deck against it.
The bigger problem is the baseline. ALM Corp’s 2026 data flips the story entirely: it argues ChatGPT converts roughly 31% higher than non-branded organic. Same ChatGPT, one source saying it’s 13% below organic, another saying 31% above. The whole gap lives in which slice of organic you compare against. The first likely folds in branded-search traffic, which is your most loyal customers and converts high by nature. The second isolates non-branded. So when you see any headline about AI converting some percent higher or lower, the first question should be: against what baseline? A conversion comparison that hides its baseline is barely a comparison.
My read: AI shopping today is high-intent and low-volume. The people who arrive are ready to buy. There just aren’t many of them yet.
What each channel costs you to join, and what to claim cheaply
Set volume aside and look only at the effort and money it takes to participate. The four diverge sharply.
Perplexity is the best cheap foothold by a mile. Free, a five-minute signup, zero fees, and they eat the shipping cost. The traffic is small, but you pay almost nothing for the slot, so there’s no reason not to take it. When a channel costs a signature to occupy, occupy it and move on.
ChatGPT is worth acting on now precisely because that 50K-250K USD ad floor came down in May 2026. Small sellers couldn’t get in before; now you can test at zero minimum spend. Paired with the new Conversions API and pixel attribution, you can run a tiny budget and pull real ROAS numbers before deciding whether to scale. I’d test this window sooner rather than later.
Google AI Mode doesn’t ask for a separate “AI move,” but it pushes the cost onto your Merchant Center feed. If your Google Shopping feed is already clean and fully attributed, AI Mode is a free passenger. If it’s rough, do the work, but file it under your SEO and feed budget rather than booking it as a standalone AI initiative.
Copilot is near-zero friction for Shopify sellers thanks to auto-enrollment; your only job is making sure you didn’t accidentally opt out. If you’re not on Shopify, it can wait.
A right-sized investment call, no cheerleading
Put it together and the operating rule is one sentence: size your resources to the 0.2% reality, but claim the free and now-floorless positions early.
Concretely. First, don’t touch your email, affiliate, or organic budget. Those three still drive the real revenue, and an 86% affiliate conversion edge plus a 13% organic edge are not noise. Moving money out of them while AI sits under 0.2% means trading certain business for a curve that’s still climbing. That’s a bad trade.
Second, claim every near-free slot. Sign up for Perplexity, confirm your Copilot enrollment, and run a small test budget on ChatGPT while the floor is gone. This costs a few hours of someone’s time, not a budget reallocation.
Third, treat feeds and citation quality as a long compounding bet. AI shopping is small today, but the direction is clear and the intent quality is high. Fix your product-feed attributes now and write content that’s easy to cite, and when the curve finally lifts you’re already on the field instead of scrambling on. Match the spend to the 0.2% truth, but point the posture at the compounding curve.
Related Articles
AI Ad Attribution Tools Compared 2026: Triple Whale vs Northbeam vs LayerFive
Meta says 100 conversions. Shopify shows 65 orders. That gap — attribution inflation — is what all three tools claim to fix. Triple Whale reconciles through Shopify pixel, Northbeam via a clean room that credits video views, LayerFive through transaction-level matching. Here is how to choose.
The Sub-$100 SEO Toolkit for a Solo Store 2026: Don't Default to Paying for Ahrefs
Ahrefs vs Semrush is the wrong question for a one-person store — you'll never touch 90% of either. This compares four sub-$100 tools — SE Ranking, Keysearch, Mangools, LowFruits — and matches each to a job: low-competition keyword mining, all-in-one, beginner UI, or built-in AI writing.
Google AI Mode Now Has Shopping Ads: What Cross-Border Sellers Need to Know
Google's AI Mode has 75 million daily users, and shopping ads are now appearing inside AI conversations. These aren't the old product carousels — they're dynamically generated recommendations woven into AI answers. Here's how to prepare.