Week 40 - The Agentic Commerce Reality Check: Google's AI Mode Click Problem, OpenAI's Checkout Expansion, and Meijer's European Tech Bet

MaxWelcome to the Checkout Point, your quick dive into this week's e-commerce buzz with me, Max, powered by Blicket. These twenty minutes pack the latest trends, news, and insights. Let's navigate the digital marketplace together. Ready? Let's go.

RachelOkay, okay, hang on. You know how everyone's been saying robots are gonna do our shopping for us now?

Speaker 3The AI agent thing? Yeah.

RachelIt's happening. Like, actually happening. I'm Max. This is Rachel. Welcome back to the Checkout Point.

Speaker 3And I'm already skeptical.

RachelYou're always skeptical.

Speaker 3It's my whole personality.

RachelFair, but listen. Everybody's racing to build the thing where an AI just buys stuff for you. No typing, no scrolling, just, "Hey, get me new running shoes," and boom, done.

Speaker 3Sounds great in a demo video.

RachelRight? That's the pitch.

Speaker 3But here's my question. Does it actually work, like, for a real person buying a real thing right now?

RachelThat's a fair question.

Speaker 3Because I've seen a lot of, "This changes everything," tech that quietly changes nothing.

RachelWow, way to kill the vibe in the first two minutes.

Speaker 3It's a service.

RachelOkay, service noted, complaint filed. But seriously, that's the tension we're chasing all episode, hype versus whether it actually delivers for real people.

Speaker 3Which, spoiler, is basically my favorite kind of episode to do.

RachelNoted. We're gonna get into whether shoppers are actually clicking through these AI experiences, what it means if merchants get skipped entirely, and there's some retail and lending news later that ties right back into this same problem.

Speaker 3The AI-shops-for-you dream versus the AI-shops-for-you reality.

RachelExactly that gap. All right, let's start with a Land Grab everyone's talking about.

Speaker 3The one where every company suddenly has an agent?

RachelThat one. So Google didn't just tease this, they actually shipped something. TechCrunch reported that Google rolled out a Universal Commerce Protocol back in January, right alongside Gemini Enterprise for customer experience.

Speaker 3Wait, so this is the actual plumbing, not just a demo video?

RachelReal plumbing. The whole point is letting an AI agent complete a purchase for you instead of you clicking through five tabs.

Speaker 3Okay, that's a real shift, not a slide deck.

RachelAnd here's the number that got me. That same TechCrunch piece cites Adobe data showing AI-driven traffic to retailer sites grew six hundred and ninety-three point four percent during the last holiday season.

Speaker 3Wait, six hundred and ninety-three point four percent?

RachelYeah. Picture your normal holiday bump, then multiply it by seven.

Speaker 3That's not a bump. That's a different graph entirely.

RachelRight? I saw that stat and just sat there for a minute.

Speaker 3Okay, so Google's building the rails. Who else is in this race?

RachelAI. Retail Dive covered this. They debuted Instant Checkout for Etsy sellers, and they're rolling it out to over a million Shopify merchants.

Speaker 3A million? That's not a pilot, that's a launch.

RachelAnd Etsy's product chief basically said the whole point is cutting shopper friction. You find the thing inside the chat, you buy the thing inside the chat. Done.

Speaker 3No tab switching.

RachelNone. You never leave the conversation.

Speaker 3I mean, that's the dream everyone in e-commerce has chased for a decade, fewer steps between want and bought.

RachelIt is, though I'll say merchants I talked to were nervous about it too. If the whole purchase happens inside somebody else's chat window, you lose a lot of the relationship. No cart abandonment email, no retargeting pixel, none of that.

Speaker 3Right. You're handing over the entire customer relationship to whoever owns the chat interface.

RachelWhich is a trade a lot of businesses might not think through until it's already happened.

Speaker 3That's a real cost people aren't pricing in yet.

RachelTotally. And Google's not sitting still either. They've been rolling out new merchant center data attributes specifically built for AI mode and Gemini to actually find your product data.

Speaker 3So it's not just showing up in search, it's showing up inside the assistant.

RachelExactly. And they're piloting something called Direct Offers, ads that let retailers put exclusive discounts directly inside AI mode.

Speaker 3Huh. So you could get an AI-only coupon.

RachelKind of, yeah. It's a whole new inventory of attention nobody's fought over yet.

Speaker 3Okay, but real talk, who actually gets access to that inventory? Because exclusive discounts inside AI mode sounds like it's gonna favor whoever's got the biggest ad budget already.

RachelProbably, yeah. I mean, that's kind of always the pattern with new ad inventory. The big spenders get first crack, and then eventually it trickles down to everyone else.

Speaker 3So a small shop probably isn't getting a Direct Offer slot on day one.

RachelRealistically, no. Early access almost always goes to whoever's already got the biggest merchant center account and the ad spend to match.

Speaker 3Which kind of undercuts the everyone gets discovered by AI pitch a little.

RachelA little, sure. Though, to be fair, that's true of basically every new ad surface that's ever launched. Google Shopping was the same story way back.

Speaker 3Fair point. New doesn't mean equal.

RachelRight. It just means new. And yeah, I'd bet the reporting on this in six months looks different once smaller merchants actually try to get a slot.

Speaker 3Okay, before you get too hyped, where's this four thousand seven hundred percent number coming from? Because I've seen that floating.

RachelOh, this is the one. There's a 2026 guide that pulls this together, and they cite an IBM study finding forty-five percent of consumers already use AI somewhere in their buying journey.

Speaker 3Almost half.

RachelAlmost half. And separately, they point to Adobe data showing AI-driven retail traffic up four thousand seven hundred percent year over year.

Speaker 3Hold on. Is that the same stat as the six hundred and ninety-three percent from a second ago?

RachelNo, different measure. That one was holiday season specific. This is a full year comparison. Different window, both huge.

Speaker 3Okay, that actually matters because people mix those up constantly.

RachelThey do. But look, either number you pick, this is the fastest I've seen e-commerce discovery move in the years I've been covering it. Like faster than mobile checkout, faster than social commerce.

Speaker 3You say that about a lot of things.

RachelOkay, fair, but this one's got the receipts. Universal Commerce Protocol, instant checkout at scale, Merchant Center rebuilding itself for AI. That's three different companies moving at once.

Speaker 3It is a lot of infrastructure going up fast.

RachelIt's genuinely wild watching three separate companies race to build the same on-ramp.

Speaker 3I'll give you that it's a lot of engineering effort pointed at the same problem all at once.

RachelRight. And usually, when three companies converge on the same idea within months of each other, that means the underlying shift is real, even if the execution's still messy.

Speaker 3Or it means everybody's terrified of getting left out and racing on vibes.

RachelThat's also a very real possibility, sure.

Speaker 3I've seen that movie before. Everybody rushes to build the same feature, and six months later, half of them quietly shut it down.

RachelSure. But even in that scenario, the infrastructure usually sticks around in some form. Somebody ends up owning the rails even if the also-rans drop off.

Speaker 3Okay, but here's the thing: a chart going up and to the right doesn't tell me anybody's actually happy using this stuff. Growth in traffic isn't the same as it working for the person on the other end.

RachelFair.

Speaker 3So before we crown anybody, let's actually check what's happening when a real shopper opens one of these AI experiences and tries to buy something.

RachelOkay. What'd they find?

Speaker 3So Search Engine Journal put out a piece this week that kind of stops the party. Roger Monttis byline published just a few days ago.

RachelOkay. What'd they find?

Speaker 3They tracked over a thousand real users going through Google's AI mode.

RachelA thousand people, not a lab test with five interns. I like that already.

Speaker 3Right. And the headline finding is simple: AI mode sends fewer clicks out to actual websites than regular search does.

RachelWait, fewer? After everything we just said about traffic exploding?

Speaker 3Fewer clicks out, and on top of that, people using it reported a worse experience.

RachelWorse how? Like slower, wrong answers, what?

Speaker 3The piece frames it as a straight up user experience problem. People don't love getting a summary instead of getting to click through and actually buy something.

RachelHuh. So Google's telling merchants, "Hey, this is the future. Plug in your inventory."

Speaker 3And meanwhile, the actual humans on the other end are basically getting stuck in the waiting room.

RachelI was gonna say the same thing, weirdly.

Speaker 3Great minds.

RachelSo how do you square that with everything from the last segment, all that traffic growth we just walked through?

Speaker 3That's exactly my problem with it. Growth in visits to a page doesn't mean the visit did anything for anybody.

RachelYeah. More people showing up to a store that keeps the door half shut.

Speaker 3Kind of, yeah. You can have huge referral numbers and still have a broken funnel underneath it.

RachelI mean, I've seen this in CRO stuff forever. Traffic's the vanity metric. Conversion's the actual scoreboard.

Speaker 3Right. And nobody in this research is saying the AI traffic doesn't exist. They're saying what happens once it lands is the part that's rough.

RachelSo it's not really an AI-is-fake problem, it's an AI-is-early problem. Maybe.

Speaker 3Or it's a Google-specific problem. Or it's a Google specific problem. The study's about AI mode specifically, not every agent out there.

RachelFair. Fair. Still, though, that's kind of a big gap between the pitch and what a thousand real people actually experienced.

Speaker 3And when you actually dig into what poor experience means here, it's not some vague vibe. It's people typing in a real search, getting handed a paragraph summary, and never getting a link to actually go buy the thing.

RachelSo the buying intent was right there, and it just stalled out.

Speaker 3Stalled out in a box on the results page, yeah.

RachelMan, that's such a nightmare from the CRO side. You've got warm traffic with actual intent, and the funnel just quietly swallows it before it even starts.

Speaker 3And it's the most expensive kind of leak because you don't even see it in your own analytics. The person never reached your site to begin with.

RachelOh, that's rough. You don't even get the data to diagnose the problem.

Speaker 3Right. You can't run a heat map on a page nobody ever landed on.

RachelOkay, that one actually stings a little.

Speaker 3It should.

RachelYeah, no kidding.

Speaker 3It is. And it matters because every retailer listening to this show has been told for a year now, "Get your feed ready. AI shoppers are coming."

RachelAnd now there's a study saying, "Sure, they're coming, but they might bounce right off you."

Speaker 3Or off Google technically. The click just doesn't happen at all.

RachelWhich for a merchant is the same outcome. No click, no sale. Who cares whose fault it is?

Speaker 3That's the retailer version of-- Actually, forget it. That joke wasn't going anywhere.

RachelWhat was that?

Speaker 3Nothing. I tripped over my own sentence. Anyway, the point stands. It doesn't matter to a retail owner why the click didn't happen.

RachelNo, totally. A missed sale is a missed sale.

Speaker 3So back at the top of the show, we asked whether an agent shopping for you is actually better than doing it yourself.

RachelAnd this is basically the first hard data point that says, "Eh, not yet. Not really."

Speaker 3Not across the board, no. It's one study, one product, but it's a real crack in the story everybody's telling.

RachelSo the hype and the lived experience are pulling in two different directions right now.

Speaker 3Which is worth sitting with because none of the numbers we talked about earlier tell you whether the person on the other end was happy.

RachelRight. Growth numbers don't have a feelings column.

Speaker 3They really don't.

RachelOkay, so if I'm a merchant listening to this right now kind of freaking out, what do I actually do with that? Just wait it out?

Speaker 3No, I don't think waiting is the move. I think you stop assuming a rising traffic number automatically means a rising sales number, and you start asking your own data that question directly. Because honestly, most people haven't even glanced at their own numbers before repeating the industry-wide stat.

RachelLike literally go check whether AI referral traffic on your own site is converting the same as regular search traffic.

Speaker 3Exactly that. Because right now, everyone's citing the industry-wide stat, and almost nobody's checked their own funnel.

RachelWhats that look like in practice, though? Like an afternoon with a spreadsheet?

Speaker 3Pretty much. Segment your traffic sources, compare conversion rate on the AI-referred visits against everything else, and see if theres a real gap. It's not glamorous, but it answers the question your gut can't.

RachelWhich, again, incredibly on brand for you.

Speaker 3I contain multitudes.

RachelMultitudes of spreadsheets.

Speaker 3Always.

RachelOkay, so if the click experience is this shaky, what's a business actually supposed to do to show up right inside these AI systems and get trusted when it counts?

Speaker 3That's the real question on the table now.

RachelBecause just having a feed isn't the same as being findable.

Speaker 3Or trustworthy. Those are two different problems.

RachelTwo different problems, one segment. Let's go figure out what actually works. Okay, so I went digging on this, and there's a piece from Search Engine Journal that actually gets into the plumbing.

Speaker 3Plumbing, meaning what?

RachelMeaning the actual stuff a business has to fix. Darren Shaw over at Whitespark and Russ Jeffrey from Duda basically argue that showing up in AI answers starts with three boring things: Can the crawler even get in? Is your business data accurate? And can a language model actually read the page without choking on it?

Speaker 3Choking on it how?

RachelLike bad HTML, weird JavaScript rendering, stuff buried behind menus. If a bot can't parse it, it doesn't matter how good your product is.

Speaker 3Give me a for instance.

RachelSure. Picture a local shop with a slick-looking site, but their hours and inventory only load in through some JavaScript widget the crawler never actually runs. A human sees it fine. The bot just sees a blank section.

Speaker 3So the page can look perfect to a person and still be invisible to the thing that's supposed to find it.

RachelExactly. Invisible and unreadable at the same time.

Speaker 3So it's less optimize for AI and more stop blocking the front door.

RachelKind of, yeah. Shaw and Jeffery are basically saying local visibility lives or dies on those three things: access, accuracy, readability.

Speaker 3Accuracy meaning exactly what? Like your hours are right?

RachelHours, address, phone number, inventory if you've got it listed. Small stuff that everybody treats as an afterthought.

Speaker 3Right, because nothing says trustworthy business like a Google listing that says you're open on a holiday when you're not.

RachelExactly that.

Speaker 3Okay, I buy the crawler thing, but is this really new? Feels like that's just SEO.

RachelIt's SEO with higher stakes because now there's a separate piece out there arguing this whole shift is turning e-commerce into a straight up data quality contest.

Speaker 3A contest?

RachelYeah. The framing is that the businesses with the cleanest, most structured product data are the ones the agent actually surfaces. Everyone else just doesn't get picked.

Speaker 3That tracks with what we've been saying all segment. The agent isn't going to guess what you sell.

RachelNo, it's gonna read what's there and move on if it's a mess.

Speaker 3So merchants basically need a data janitor now.

RachelKind of. Titles are shifting because of this stuff.

Speaker 3Okay, so here's my actual concern with all of this. Merchants can clean up their data forever and still have zero idea if it's working. Is there any visibility into that?

RachelThere's something coming. Google's Merchant Center help documentation describes new AI performance insights still labeled Coming Soon that'll show brand visibility and how you're performing across the AI mode and AI Overviews funnel, plus the Gemini app.

Speaker 3Wait, a dashboard for how you show up inside the AI itself?

RachelThat's the pitch. Basically, traffic and conversion, but for the AI surfaces instead of regular search.

Speaker 3See, that's the piece I actually care about because right now merchants are flying blind. You could fix your crawler access, clean your feed, do everything Sean Jeffrey are describing, and have no way to know if any of it moved a single sale.

RachelRight, and that's the gap.

Speaker 3So until that dashboard's actually live, my advice to any merchant listening is don't chase the shiny agentic thing. Fix the boring thing. Crawler access, accurate data, those are the two levers you can pull today that have a real measurable payoff even before Google gives you a report card.

RachelYou're basically saying skip the hype, do the homework.

Speaker 3I'm saying track it like you'd track any other channel. If you can't measure it yet, at least make sure the inputs are clean because when the measurement does show up, you don't wanna be starting from zero.

RachelThat's a very Rachel answer.

Speaker 3I contain multitudes.

RachelMultitudes of spreadsheets.

Speaker 3Always.

RachelAll right, so that's the fix-it list: access, accuracy, readability, and a scoreboard that's on the way.

Speaker 3Yeah, that's plenty on the AI side for one week. There's actually a bunch of other stuff that happened in retail and fintech that we haven't touched at all.

RachelOh, right. The roundup.

Speaker 3Let's get into who else made news this week: grocery, lending, cross-border, all of it.

RachelOkay, first one, Meijer, the Midwest grocery chain.

Speaker 3Oh, I know Meijer. Big box, groceries, everything under one roof. Right, and they're teaming up with a European company called StrongPoint to speed up how they pick online grocery orders.

RachelWait, European? Like actually based in Europe?

Speaker 3Yeah. Progressive Grocer had this one. StrongPoint's tech is basically built for warehouse and store picking efficiency.

RachelSo somebody's in the back room right now running around with a cart filling your pickup order.

Speaker 3Exactly, and if that process is slow, your order's late or wrong, and you don't order online again. That's such a boring problem compared to AI shopping agents, but it's the actual business.

RachelIt's the whole business. Nobody writes a headline about picking accuracy, but it's most of the cost. Yeah, and honestly, that's the same principle as an AI agent doing your shopping too, right? Nobody sees the picking accuracy or the underlying plumbing. They just notice when their order shows up wrong.

Speaker 3Exactly. The invisible stuff is what actually keeps customers coming back or scares them off. It's not glamorous, but it's the whole game.

RachelWhich I guess is the theme of basically this entire episode, now that I think about it.

Speaker 3Kind of, yeah. The flashy front end gets the headlines. The boring back end decides whether the business survives.

RachelNo kidding. Okay, next. This one's fintech, Kastle.

Speaker 3Kastle raised a Series A, twenty-four million dollars.

RachelTwenty-four million? For what exactly?

Speaker 3They're an agentic AI platform for consumer lenders, so banks and lenders using AI agents to handle loan decisions, underwriting, that kind of work.

RachelSo it's the same agentic wave we've been talking about all episode, just showing up in lending instead of shopping carts.

Speaker 3Right, and PYMNTS reported the round closed the week before they wrote it up.

RachelTwenty-four million isn't nothing, but it's not exactly a mega round either.

Speaker 3No, it's an early bet. Investors think lenders are going to need this plumbing the same way retailers do.

RachelDo we know anything about who's actually backing this round, or is that not public yet?

Speaker 3The reporting didn't name the whole syndicate, just that it was a Series A. Early stage stuff like this usually stays a little vague until there's more traction to show.

RachelMakes sense. Twenty-four million buys you a runway, not a victory lap.

Speaker 3Right. This is very much a let's see if lenders actually adopt this stage, not a we've already won stage.

RachelEverybody wants to build the pipes before anyone's sure the water is even running.

Speaker 3Which, funny, brings us right back to the theme.

RachelOkay, last one. This is more macro. TechRepublic wrote about small businesses across Southeast Asia.

Speaker 3ASEAN countries. Small and medium businesses there are going digital fast.

RachelWhich sounds great on its face.

Speaker 3It is, but the piece is pretty clear that cross-border e-commerce in that region is still a mess. Marketplace access, payments, logistics, regulation, all of it, still friction.

RachelSo you can have a store online and still not actually be able to sell into the next country over.

Speaker 3Right. Going digital doesn't mean the border problems disappear. Somebody still has to move the box and clear customs.

RachelI mean, come on. That's true everywhere though, right? Digital doesn't erase geography.

Speaker 3No, it just makes the geography more visible, honestly, because now the customer expects it to be instant.

RachelThat's rough because customer patience for cross-border shipping was already thin before AI-everything made every other part of shopping feel instant too.

Speaker 3Right. Expectations went up everywhere else, but customs and logistics didn't get any faster just because checkout did.

RachelYeah. Okay, so three totally different stories: a grocer fixing warehouse picking, a lender funding AI underwriting, and small businesses hitting the same old cross-border walls.

Speaker 3But they've all got the same thread if you squint. Nobody in any of these stories is chasing a flashy AI shopping demo.

RachelThey're fixing the actual operational stuff: picking speed, underwriting speed, getting a package across a border.

Speaker 3Meanwhile, the headlines this week were all about agents doing your shopping for you.

RachelRight, and the real work is happening kinda quietly underneath all that.

Speaker 3Efficiency and trust built on the ground while the big flashy stuff gets all the attention up top.

RachelWhich honestly tracks with everything else we talked about today.

Speaker 3It really does.

RachelAll right, that's a wrap on the news. Let's bring it home.

Speaker 3So if you're running a store and wondering what to do with all this agent stuff, we already gave you the specific fixes last segment: crawler access, accurate data, all of it.

RachelRight.

Speaker 3The tech is racing ahead, and the trust piece is lagging behind it. That gap is where merchants either close ground or fall further behind.

RachelStill the unglamorous stuff, though.

Speaker 3Unglamorous stuff that pays off whether an AI agent ever shows up or not. That's the part people keep skipping.

RachelBecause everyone wants to talk about the robot doing the shopping.

Speaker 3And almost nobody wants to talk about whether their own product data is even correct.

RachelWhich, let's be honest, is way less fun to tweet about.

Speaker 3It is, but margins don't care what's fun.

RachelFair.

Speaker 3If there's one thing from this whole episode, it's that hype moves faster than trust does. Close that gap, and you're ahead of basically everyone else still chasing the shiny thing.

RachelOkay, I'll take the easy job now.

Speaker 3Go for it.

RachelThanks for hanging out with us today, seriously. If you like the show, follow us, leave a rating, tell a friend who's still arguing about whether AI is gonna kill retail.

Speaker 3There's always one.

RachelThere's always one. We'll see you next week with more of this stuff.

Speaker 3Take care, everybody. And hey, go check your own analytics before next week. You might learn something.

RachelHomework on a podcast. Classic Rachel.

Speaker 3Always.

RachelPeace.