Week 15 - Shopify’s AI playbook and agentic shopping: fuel surcharges, WTO deal, and new rules for cross‑border sellers

Max: Welcome to the Checkout Point, your quick dive into this week's e-commerce buzz with me, Alex, powered by Blicket. These 20 minutes pack the latest trends, news and insights. Let's navigate the digital marketplace together. Ready? Let's go! Rachel: What is up everybody, welcome back to the Checkout Point. Speaker 3: Hey, good to be here. If you're running an e-commerce brand, this episode is loaded. Rachel: Oh yeah, Because Shopify kind of woke up one day and said, Surprise, we're an AI company now. Speaker 3: Right. Selling straight inside ChatGPT, running these cheaper models like Qwen 3 to crunch all your store data, that is a very specific kind of flex. Rachel: And here's the thing the AI stuff is cool, but it's also how they dodge all the rising fees everywhere else in the stack You know what I mean? Speaker 3: Exactly. Because while that's happening, Amazon quietly drops this 3.5% fuel and logistics surcharge. That number sounds tiny until you look at your per unit margin and want to scream. Rachel: Wait, really some of you are gonna lose like a dollar a unit and not notice until Q4 Or we're walking through how to claw that back with boring but powerful stuff like resizing boxes and mixing FBA with 3PLs. Speaker 3: And a few sneaky price nudges that don't crush your conversion. I've got a story on that where one tiny change fixed up why are we even doing this SKU. Rachel: The why do we sell this at all report is my favorite dashboard. Speaker 3: It really is. Rachel: So anyway, once we plug the leaks, we're going into agentic. authentic commerce, like full AI shopper doing the browsing, comparing and checking policies for you. Speaker 3: Day in the lifestyle: You wake up, your agent has already picked the product, checked shipping, handled taxes, but only if your catalog is clean, structured, and not a hot mess. Rachel: If your data is chaos, that agent is never picking you. We'll hit what merchant-ready actually looks like in plain English. Speaker 3: Yeah. Then we zoom out on global rules, WTO stuff, and how cross-border AI AI shopping changes pricing, duties, and compliance. I promise, more practical than it sounds. Rachel: Totally. And we've got a quick blog highlight on AI-driven margin wins, so stick around for that. Speaker 3: Alright, let's get into it. First up, Shopify's quiet shift into AI and what it means for your revenue and costs. Rachel: Grab your numbers, grab your coffee, we're hitting that first segment right now. Three easy ways to get your questions to us. Check the description for our web form, text us at 747-293-4612, or call the same number to leave a voice question. However you reach out, we can't wait to hear from you. Okay, quick hot take to start. Shopify is quietly turning into an AI company. Speaker 3: Totally. And not in a hypey, we added a chatbot way. This is like actual money stuff for merchants. Rachel: Yeah, the headline for me is this. You can literally sell inside ChatGPT now. Your Shopify store shows up right in the chat. Speaker 3: Right. So picture a skincare brand. A shopper types, my skin is dry and sensitive. what should I use in ChatGPT? Rachel: And instead of just vibes and blog posts, ChatGPT can pull your real products, live prices, inventory and say, here are three options from this store, then complete checkout in that chat. Speaker 3: No pop-up, no new tab, no cart steps. Just, cool, I'll take the second one and done. Rachel: That is wild, right? You basically get a sales rep sitting inside whatever AI chat people are already using. Speaker 3: And here's the key thing merchants care about. If you're not doing paid ads well, this can feel like free incremental revenue. You show up where the intent is. Rachel: Now the obvious question is, how is Shopify paying for all this AI magic? Because those models aren't cheap. Speaker 3: Yeah, so this Qwen3 thing, let me put that in simple dollars and cents terms. Rachel: Please, because once we say model family, people tune out. Speaker 3: Exactly. So Shopify tested this new model. model family, called Qwen3 for workloads like reading product feeds and Rachel: Wow. Speaker 3: support tickets. Rachel: They found it could be around Seventy times cheaper to run for those tasks. Speaker 3: Wait, really? Seventy five? Rachel: Yep. So imagine you were paying $75 for a chunk of AI processing. Now you're paying one. Speaker 3: That's a clearance sale. Rachel: And what that means is they can run AI on way more merchant data without the bill exploding. Think of scraping all your product descriptions, tags, reviews, and orders non-stop. Speaker 3: So instead of we can tag 10 products and then cry over the AWS bill, it's we can tag the whole catalog every day. Rachel: Exactly. Or auto-cleaning titles, pulling attributes, spotting which reviews mention sizing, all that boring admin stuff. Speaker 3: Boring, but it prints time. Rachel: And time is margin. Speaker 3: So we have AI-native selling channels like ChatGPT. and then cheap enough models to chew through all the data in the background. Rachel: This is where their whole AI power stack story comes in. They're not just shipping one feature, they're building the plumbing. Yeah, they're basically saying every team inside Shopify now asks, can AI do the grunt work here first? That is the default. Speaker 3: Let us make that concrete. Take a small apparel brand. Today, you install a few apps and suddenly your search bar actually understands black dress for a winter wedding instead of just the word "black. Rachel: Your recommendations stop being random; the AI has seen your orders and knows people who buy boots often add the waterproof spray. Speaker 3: So average order value creeps up a few dollars because the complete the look box is finally smart. Rachel: And for conversion, think of on site chat. But instead of a clunky bot, it has your policies, your catalog, your sizing charts in its head because Qwen three chewed through everything. Speaker 3: So a shopper types "Does this jacket run small?" and the chat can answer "Most customers say it runs half a size small, consider sizing up," then drop a one click add to cart. Rachel: That's the stuff Shopify likes: quiet lips. One percent here, three percent there. Speaker 3: And you, as the merchant, don't have to become an AI engineer, you just flip on apps that sit on top of Shopify's AI stack. Rachel: I mean, most brands are fine with that trade, if revenue is up and support tickets are down. Speaker 4: down! Speaker 3: Sure—as long as the savings don't get wiped somewhere else on the P and L. Rachel: That's the tension I keep coming back to: AI might save you money on data, but surcharges can eat it right back. Speaker 3: So the real question is, while your software stack is getting smarter and cheaper, what happens when your logistics partners quietly nudge their fees up a few percent? Max: Shifting gears for a second, we have to talk about Amazon's new Fuel and Logistics thing. Yeah, the fun part, fees. Rachel: They're adding a 3.5% surcharge on top of certain fulfillment and logistics fees. Think FBA and some of the related shipping services. Max: So just to be clear, that's not 3.5% of your sale price, it's 3.5% on the fees you already pay them, right? Rachel: Exactly. It stacks on what you're already getting billed. And Amazon is pointing to energy costs and the chaos around oil because of the... Because of the war around Iran and the region. Max: Cool. Love when geopolitics hits your P&L. Rachel: Yeah, and it sounds tiny, 3.5, who cares? But on thin margins, it bites. Max: Okay, do the scary math for us. Rachel: Alright, super simple example. Say you sell a product for $30. Your landed cost, so product plus freight into Amazon, is $12. Max: Okay, 30 revenue, 12 cost. Got it. Rachel: Your normal FBA fee plus referral fee is. Say that all in is ten dollars. I'm simplifying, but roll with it. Max: So 30 minus 12 minus 10, you're left with $8 profit per unit. Rachel: Right. Now Amazon slaps 3.5 percent on that ten dollar fee. That's 35 cents. Max: Wait, so you go from eight profit to 7.65? Rachel: Yep, you just lost 35 cents on every unit without changing anything else. Max: That's like, what, a little over 4% of your profit gone? Rachel: Roughly, yeah. Yeah, and if you were only making $4 a unit, that same 35 cents is almost 9% of your profit. Max: Ouch. And that's just one tweak on one line item. Rachel: And remember, for some categories the base fee number is higher, so the absolute hit is bigger. Max: So if you're doing, say, 5,000 units a month, that tiny surcharge is more than $1,700 a month out the door. Rachel: Exactly. Real money. Max: All right, now I'm stressed. What can people actually do this quarter instead of just complaining about Amazon? Amazon in a group chat. Rachel: Three fast levers. First one is size and weight. A lot of people never re-audit this. Max: Like the poly bag from 2020 that's still oversized. Rachel: Yes, or that box that's one inch too big and bumps you into the next FBA tier. If you can shrink packaging, flatten something, or bundle smart, you might drop a tier and offset the whole surcharge. Max: So practical homework there is pull your FBA fee reports, look at dimensions, and see if you're right at a cutoff. Cutoff. Rachel: Exactly—and literally measure your product yourself! People trust Amazon's listed dimensions like they're carved in stone. Speaker 3: Yeah, Max: And sometimes they're just wrong. Rachel: Totally. Second lever is pricing. I know people hate raising prices, but a tiny change can cover this. Max: Like if you're at twenty nine ninety nine and you go to thirty ninety nine, most shoppers don't blink. Rachel: Right, and that extra dollar can more than cover the thirty five cents you lost. You do not have to- to eat every fee increase. Max: But you also don't want a new conversion. Rachel: For sure. So test it. Maybe only raise on bundles or on your best sellers first. See what happens to conversion and to profit per session, not just per unit. Max: Yeah, if you drop conversion a hair but you keep more margin, you can still end up ahead. Rachel: Third lever is how you mix fulfillment. Don't assume everything needs to be FBA all the time. Max: Okay, let's dig into that because some people here leave FBA and... and panic. Rachel: I'm not saying abandon it. I'm saying split. Maybe your fast movers stay on FBA because the Prime badge is gold, but your slow, bulky stuff lives in a 3PL or your own warehouse and you use Seller-Fulfilled Prime or just standard shipping. Max: So you basically keep Amazon for the SKUs where the conversion boost pays for the higher fees. Rachel: Yes, and move the awkward SKUs that get hammered by storage and surcharges. is off to cheaper options. Max: What about non-Amazon channels in that mix? Rachel: Huge. If you're sending the same inventory to a 3PL that can serve your own site, Walmart, whatever, that freight and storage gets spread across more revenue streams. Max: So even if Amazon skims more on one side, you're making it back with better margin on your direct site. Rachel: Exactly. And this is where knowing your real unit economics matters. Not vibes. Actual numbers. Max: Vibes per unit is not a KPI. Rachel: Sadly. Okay, rapid checklist for people listening: one, check your FBA dimensions and weight and see if packaging tweaks could drop a tier. Max: (Two) test a small price bump on key SKUs and watch profit per session not just unit margin. Rachel: Three, look at your product catalogue and flag the weirdos—bulky, slow, seasonal. Ask: Do these really need to be in FBA? Max: Exactly, and if you're already investing in smart tools and AI to earn more on the top line, you want that money to survive the fee gauntlet, which nicely points to where shopping is going: because if these costs keep climbing, the only way to stay sane is to upgrade how people buy, not just how you ship. Rachel: Yeah, like giving every shopper their own little digital rep who only shows them things that actually fit, budget wise and preference wise. Max: We're talking about that whole agentic commerce idea, Right? Where an AI agent shops with or for your customer. Speaker 4: Exactly. Because if fees are creeping up, you need smarter journeys so every visit throws off more profit. Max: Alright, let's get into how those AI shopping agents actually work and what ready for agents looks like for your store. Shifting gears a bit, let's talk about shoppers basically having their own robot buyer. Speaker 4: Their own little shopping intern on the payroll. Max: Exactly. So simple definition. Agentic commerce is when an AI agent doesn't just recommend, it actually goes out, compares options, and can buy stuff for you within rules you set. Speaker 4: So like instead of me opening five tabs, this thing opens 5,000 in the background and brings back here's the best one. one. Max: Yeah, Picture this: you tell your agent "Kid sneakers under sixty bucks, wide feet, no neon, needs to arrive before Saturday soccer." Then you go do life. The agent checks brands, shipping cutoffs, return policies and just handles it. Rachel: And it actually remembers the wide feet instead of you yelling at your laptop every three months. Max: Right. Or travel. You say "work trip, two nights, near the conference center, hotel that won't roast my company card. Card. Agent finds room, checks your loyalty stuff, reads the reviews for noise complaints, then books. Speaker 4: Okay, so that sounds great for shoppers, but for merchants listening, the translation is, you are suddenly selling to a robot that is picky. Max: Yeah, this thing is like the nerdy personal assistant who reads every fine print. Speaker 4: Which is where Merchant Ready comes in. If your data's messy, that assistant just skips you. It never even shows your product to the human. Max: So break that down. Down! What does 'Merchant Ready' mean in plain English? Speaker 4: Three buckets-clean product data, structured catalogs, clear policies. Max: Short and painful. Speaker 4: Very painful. Let me translate, though: clean product data, titles that say what the thing is, solid descriptions, filled in fields. Max: Looks like size, material, color. Rachel: Not Cozy Vibes Set, and that's it. Max: Exactly; the agent has no clue what that is-is it pajamas, Candles, a cult? It needs Women's Cotton Pajama Set, long sleeve, navy, size range XL to 3XL. Rachel: Slightly less culty. Max: Second piece: Structured catalogue; so proper collections, tags used consistently, variants set up correctly; if your size is typed Small on one product and S on another the agent may think you do not have the size your shopper wants. Rachel: Oof! Max: Yeah. Last piece-clear policies. Shipping times, fees, returns, warranty-all spelled out in text the AI can read, not buried in an image of your cute policy card. Rachel: So if my return window is only in a PDF from twenty seventeen, I'm basically invisible to the agent. Max: Or worse, it guesses wrong about you, and routes the shopper to a competitor. competitor that looks safer. Rachel: Okay, so agents are like super picky comparison shoppers. How do platforms like Parallel on Shopify fit into this? Max: Right. Parallel is a good example. It sits between shoppers and merchants as this agent layer. A shopper can say, handle my birthday gifting, and Parallel's agents go hunt inside Shopify stores that are compatible. Rachel: So if you're not agent compatible, you're missing that traffic. Traffic. Max: Exactly. You miss traffic you didn't even know existed. Rachel: And that traffic is high intent because if someone trusts an agent enough to say, just buy it, they're not browsing for fun. Max: Yeah, this is not the I'm bored scrolling at midnight behavior. This is I need a thing, don't waste my time. Rachel: So what does getting ready for something like Parallels look like this week? Not in theory. Max: Super concrete. Step one, pick your top 10 products by revenue. Just ten. Open each one and ask: "Could an AI understand what this is and who it's for from the data, not the photos? Rachel: If the answer is "Eh?" fix those first. Max: Yep—add missing attributes, dimensions, ingredients, materials, care instructions, who it fits. Use the standard fields your platform gives you. Don't just shove everything into one long description. Rachel: And maybe clean up weird naming like "Bundle Three Blue. Blue, which means nothing. Max: Exactly; call it three pack bamboo kitchen towels blue-boring is good, boring is machine readable. Rachel: New Tag Line: "Be Boring, Get Rich. Max: Honestly a little bit. Step two, go to your shipping and returns pages and read them like a lawyer and a robot at the same time. Rachel: So no cute line like "We'll make it right, pinky promise" as the only rule. Max: You can keep the pinky promise, but add "30 day returns," "free within the U.S.," "customer pays return shipping internationally," in a clean sentence. Agents can work with that. Rachel: And that matters Max: Yeah. Rachel: because the agent is scoring you on risk. Can I get this here on time? And what happens if it's wrong? Max: Exactly. The more clear you are, the easier it is for the agent to justify picking you. Rachel: Alright, zooming out, we've got agents making buys, maybe through something like Parallel across stores, across countries. That suddenly makes the boring global rulebook way spicier. Max: Yeah, because your shopper's agent does not care about borders. It just cares about best option globally. But your tax, duties and shipping rules absolutely care. Rachel: So next we should talk about how the rulebook is changing, like if an agent in Europe auto-orders from a tiny U.S. brand, who's on the hook for digital duties, customs, all that fun stuff. Max: Exactly. Some new trade decisions are quietly making cross-border e-commerce easier in some places and trickier in others. Rachel: So stay with us. After the break, we're going from robot shoppers to the people riding the global rules that decide how far those robots can roam. Max: And what that means for you if your next best customer ends up living on the other side of the world. Rachel: Shifting gears for a sec, let's talk about the global rulebook your AI shoppers are about to run into. Max: Yeah, because once those agents start buying across borders, the boring stuff decides if the order even works. Rachel: So start with the big one, WTO e-commerce thing. What actually changed in human words? Max: Okay, simple version. A big group of countries basically shook hands and said, we're not slapping customs duties on digital stuff. STUFF. Rachel: So like no extra border tax on a Shopify app or a design file or a video course? Max: Exactly, and that matters because it keeps the internet from turning into a toll road for small exporters. Rachel: BGHOST1mhm. Right. If you sell digital goods, agents can shop you globally without mystery fees popping up at checkout. Max: And even if you sell physical stuff, that signal still helps: buyers trust cross border more when rules feel stable and costs are clear. Rachel: So basically the WTO is telling your agent: Go ahead, compare prices across borders. We're not booby-trapping the digital rails. Max: Yeah, not perfect, but directionally friendly for small brands going global. Rachel: Okay, zoom into country level because that's where people trip. India first. Max: India did two big things. One, they removed this tight value cap on what you could ship out by courier as an export. Rachel: Meaning? Max: Meaning, if you're a small brand sending orders to Indian customers from... From India you're less boxed in by "Oops! that order's too big for courier rules. Rachel: So if I'm a seller doing, say, higher ticket skin care sets, can actually send full bundles without needing some heavy weight freight set up. Max: Exactly: more items, higher basket sizes, still moving through courier-faster, simpler. Rachel: And two? Max: Two is their cleaning up e commerce export rules overall-more digital documentation, clearer codes. less weird gray areas on whether your order counts as an export. Rachel: So if I'm a seller using an AI agent to route orders, that agent can see, Yep, this is legit export compliant instead of hitting a customs black hole. Max: That's the idea: you'll still need a partner who knows Indian paperwork, but the rules themselves are lining up better with how e-commerce actually works. Rachel: Alright, jump to China. What's changing there? Max: China dropped new guidance on cross-border e-commerce. A risk that basically says we like this channel but play by the rules. Rachel: Very on brand. Max: Totally. Practically, it means clearer lists of what products can use special cross border tax rates, more focus on product safety and real enforcement on data and labeling. Rachel: Give me a Seller example. Max: Say you sell baby products into China through a cross border marketplace. The guidance pushes platforms and sellers to make sure it Sure, ingredients, labels and health claims match Chinese rules, not just your home country. Rachel: So if your AI agent thinks, this diaper cream looks cheap, but your labels don't match the allowed list, Max: At first, that order might never clear. Rachel: Exactly. The opportunity is huge, but you cannot wing it. You need local compliance dialed in. Max: Okay, so big picture. WTO keeps the pipes open for digital stuff. India makes shipping out easier. China tightens but clarifies rules for inbound. Rachel: That's the pattern. Less randomness, more written rules, which is great news for software and agents. and trying to automate all this. Max: So now moves. If I'm listening and want to lean into cross-border while this stuff aligns, what should I actually do this quarter? Rachel: First move: test one new market, not five. Pick a country where you already see some organic traffic or social interest. Max: Mike, you notice random Canadian orders showing up. Cool. Canada, you test. Rachel: Right. Localize shipping options and taxes just for that one. One, make sure your product data has HS codes, materials, and origin filled in so agents and apps can calculate duties cleanly. Max: Second move, revisit your landed pricing. What does the shopper pay end to end in that new market? Rachel: Yeah, pull a few test cards through from that country, look at product price, shipping, tax, any import fee. Max: And if the total looks silly, like double the product price... You either adjust your base price, eat some shipping, or stop targeting that market. Rachel: Exactly. Do that math before the AI agent decides you're bad value and quietly drops you from its short list. Max: Third move, and I feel like this gets ignored: document your rules where a robot can read them. Rachel: Yes, put return windows, country restrictions and warranty terms into structured fields in your platform, not just a wall of text. Text Max: So when some travel agent bot is picking a store, it can see: ships to Germany, thirty day returns, no hazmat issues. Rachel: And tie in a partner who knows your target country's import rules-could be your 3PL, could be a trade consultant-but give your ops team one adult in the room for each key market. Max: So the play right now is one new market, clean landed pricing, machine readable policies and Speaker 5: a new business model. Max: and at least one human expert on the gnarly local rules. Rachel: That's it! You don't have to become a trade lawyer; you just have to be clear enough that both humans and agents know your store is a safe yes across borders. Max: And the cool part is, as WTO stuff and national rules stabilize, every hour you spend cleaning this up pays off over more and more orders. Rachel: Yeah, think of it as future proofing your store for a world where your next Next best customer might not even be in your time zone, or human. Max: On that slightly scary note, we'll get into how to turn all of this into an actual checklist in a minute. Rachel: Yeah, we'll bring it back down from global policy to what you tap in your dashboard tomorrow. Max: So if you remember anything from today, remember that moment we called Shopify an AI company hiding in plain sight. Rachel: Yeah, Max: That was my villain origin story right there. Rachel: Yeah, that line is going to follow you, Max. But seriously, that shift matters. If your tools are getting cheaper and smarter and your fees are creeping up, every decision on margin suddenly counts. Max: Exactly. One sentence takeaway? Make AI work on your cost side before surcharges quietly steal the win. Rachel: Mhm, and, um, don't just nod at that. Pick one thing from today, whether it's cleaning product data or checking those new fees and actually do it this week. Max: Playful, and while you're in execution mode, hit subscribe on the Checkout Point, drop a quick review, and send this episode to the one friend who keeps saying "I'll fix my margins later. Rachel: We all know that friend. I mean, for me, that used to be me. So, Right, learn from my mistakes, not your next payout report. Max: Warmly. Thanks for hanging out with us. Seriously. This stuff is nerdy, but it's also the profit engine for your store. Rachel: Alright, that's it for today. Take care, keep your margins healthy, and we'll catch you at the next Checkout Point.