How to Sell in the Age of Agentic Shopping #616

There’s a lot of noise right now around AI in ecommerce. Agentic commerce. AI checkouts. Bots buying on behalf of customers. It all sounds big. It all sounds important. And it all sounds like something you’ll deal with… later.

But here’s the problem with that thinking: The biggest shift in Agentic Shopping isn’t happening at checkout. It’s happening the moment your customer opens ChatGPT and asks, “What should I buy?”

And that moment is already here.

In today's Playbook: 

• Why AI is becoming a new discovery channel 
• How product data determines whether you get recommended 
• Why AI is already influencing your revenue (without attribution) 
• What leading operators are doing differently

Connect with James
Explore Shopify
Shopify’s main episode
HealthPost’s main episode
Nutrition Warehouse’s main episode
StudioHawk’s main episode

Subscribe to the Add To Cart newsletter

Connect with Nathan Bush on LinkedIn

Join the Add To Cart Community 

Transcript

SPEAKER_02 0:01

Marketplaces in Australia aren't just emerging or a nice side piece to D2C anymore. They're established, they're complex, and they are moving fast. Amazon continues to accelerate, TikTok shop is looming, and for most brands, managing a marketplace properly has become a full-time job. Forecing, content, ads, logistics, compliance, it's a lot. And it requires specialist knowledge. That's where Patent can help. Patent don't just advise on marketplaces, they actually buy your product and then sell it on global marketplaces as your partner. That means they have real skin in the game. They only win if you win. Simple as that. Through their marketplace accelerator model, Patent handles everything from inventory forecasting and listing creation through to advertising, fulfillment, and international expansion. So if marketplaces feel overwhelming, or you know you should be doing more, but you just don't have the time, the team, or the energy, contact Patten to help you make the most of the marketplace opportunity. Learn more at au.patten.com. My goodness, isn't there a lot of noise right now about Agentic Commerce? Whether it's buying directly inside ChatGPT, Shopify's universal checkout protocol, AI agents completing transactions without a human even visiting your website. And if you've been following any of it, or some of it, it might feel like something that's coming, something to prepare for, something that will matter eventually, or it might feel like something that's on your doorstep and you're starting to panic a little bit. But regardless of where you stand on whether it's something you've got to think about in the future or something that you're already behind on, what is clear is that AI is already influencing which brands get discovered, which brands get considered, and which brands get chosen, even if it's not directly in the checkout. In the moment a customer opens Chat GPT or Claude and types in, what's the best running shoe for someone with flat feet, not looking at me, or find me a sustainable gift under$50. That research is happening right now, every day across millions of queries. And the brands showing up in those conversations aren't necessarily the ones with the biggest budgets or the most sophisticated tech stacks or the nerds doing AI. They're the ones whose product data is really clear, rich, and detailed enough for an AI to understand and recommend them. We were lucky enough to sit down with James Johnson from Shopify to talk about where all of this is heading. Shopify tends to set the tone, let's call it that, for where e-commerce is going. And what James shared around what they're already seeing in their data is very much worth paying close attention to. Let's hear it directly from James.

SPEAKER_01 2:59

So Agentic is really transforming retail in many respects. To make it as simple as possible, I think about it as an umbrella term of agents interacting with agents to buy something. We know and we've seen from Shopify data and broader market data that AI is fundamentally changing how we shop. So moving from search to conversation. And that's really moving into this world where shoppers describe what they want with a conversational agent, and those agents are servicing the right products. So, in those simple terms for lay people, it's really as simple as this being another channel to sell and engage with consumers. So historically, we've had things like social commerce, we've had a whole variety of ways to sell as retailers and as merchants. The AI agent shopping is really the newest channel to grow and is increasingly relevant. So, as I mentioned, Shopify is all about building the infrastructure so that retailers and merchants can jump directly into those AI conversations and making it possible so that you can transact directly in chat and have that really relevant experience. The important aspect as it relates to UCP and Agentic, again, think UCP is the plumbing. We've built this open standard so that others can build on top of it. As a consumer, you will be using UCP, but you won't be interacting with it in a discernible way. You'll be interacting with an agent that will allow you to purchase. And so those agents are ones that we know, like Google in AI mode or Gemini, Microsoft Copilot, ChatGPT. They will be the agents that consumers will interact with. And fundamentally, this is about Shopify syndicating your products into all of those AI channels accurately in real time.

SPEAKER_00 4:46

Okay. So Shopify UCP is making sure that your products as a retailer are there when anybody's in their private chat with ChatGPT, Claude, wherever they are, they're in there and Shopify is like giving you the ability to like present your products like where and when this customer needs them. So it's not an agent, like some people might think, like how people are paying for bots to go and shop and do their tasks for them.

SPEAKER_01 5:12

Correct. And I think the important aspect there is that merchants and retailers are in control. So you're choosing where you sell, you're choosing where you show up in those AI channels from a checkout perspective. You remain that merchant of record. So you still have views over customer data, you still have responsibilities to ship the right product to the right customer. The easiest way to think about it is this is a new channel to grow or to think more broadly from a distribution perspective.

SPEAKER_02 5:38

And do you think UCP is about getting ahead of the game? Because obviously AI is evolving so quickly, right? And we're still learning around even advertising, what's going to be possible in different channels, like where can you place your ads, what the will the ads look like? Is the idea that this gets ahead of the game, especially with the partnership between Google and Shopify, to go, no matter where you're going to sell, we want one way for retailers to set up their data to make sure they're not doing it multiple times for multiple agents.

SPEAKER_01 6:05

Yeah, the the multiple times for multiple agents is the challenge, right? So take a step back and think about the partnership. I think the really exciting thing is that Google is an expert in search and discovery. They've done that for a significant period of time. Shopify are the experts in commerce. We've been doing this for two decades now. And so it makes sense to partner to co-develop the protocol around those strengths. And so we have brought a wealth of expertise in terms of all of the challenges and quirks of checkout and commerce because it is so broad and we've seen it so so broadly over time. So that's important from a relationship perspective. In terms of as we think about setting up those merchants and those discussions, power in commerce is really hard at that scale. It means you need to have that structured product data that you mentioned. There are checkout intricacies, there's fulfillment logic, there's trust that needs to be built through all these pieces. So, one, that's that's a particular reason why we partner with Google around this. And the power of Shopify, particularly as it relates to having that information once from a product catalog and then sharing it to the right places, is a benefit for retailers and merchants. Because the other way is you're building your own custom integration with all of these platforms. And just thinking purely from the innovation and the scale perspective, we have thousands of engineers that wake up every day and build for commerce. From a retailer perspective, that's a challenge to then go, I now want to move on to whatever the next tool is from an AI perspective. So, really part of the proposition here is have all your products there, make sure they're enriched, make sure they're communicating the right information, but then syndicate that to Google, to ChatGPT, to the Shop app, to all those aspects of where your consumers are going to interact and transact with you.

SPEAKER_02 7:50

So those numbers are obviously from a small base. James was very clear about that. But when Shopify is tracking a signal that strong and that consistent, it's definitely worth asking what it means for where your business should be heading, whether you're on Shopify or not. Here are three things that operators can act on right now without waiting for any new integration or any new platform update. Number one, just go and search yourself before you do anything else. Before any strategy, before any content audit, before any conversation with a new SEO agency, do this one thing. Open Chat GPT, Clawed Perplexity, Google's Gemini, AI mode. Type in what your customer would actually say. Not a keyword, a real question. What's the best natural skincare for sensitive skin? Where can I get custom jewelry made in Australia? What running shoe is good for flat feet and long distances? See what comes back. See whether you appear, see who does. The majority of operators that I've spoken to haven't even done this really, really simple step yet. And the ones who have tend to move pretty quickly afterwards because what you find tells you more about your actual visibility than any analytics dashboard currently can. When Kelly Slessor did this exercise early on with Bard, what she found was genuinely confronting. The results were completely different to what she was finding on Google. Brands that ranked well on traditional search were not showing up in AI. Products were being recommended directly without a single link to a website. The entire discovery experience looked different. Her advice was simple. Simple step. Number two, your product data is your new storefront. Here's the thing about how AI recommends products. It can't see your photography easily. It can't feel your brand. It can't experience your store design or your packaging or the carefully crafted copy on your homepage. Or it doesn't care. What it can do is read and organize your product data and use that to decide whether you're worth recommending. Which means if your product descriptions are thin, your specifications are incomplete or your catalog is inconsistent, the AI just doesn't have enough to work with. It moves on to a brand that gives it exactly what it wants. One of the most instructive examples I've come across is with HealthPost, Abel Butler built what is probably one of the most AI-ready product catalogs in e-commerce, and he didn't do it for AI at all, kind of stumbled into it. He did it because his supplier onboarding process demanded transparent, detailed product information as a condition of them ranging products, sourcing, ethics, ingredients, production methods, all of it needed to be documented before they sold it. The AI readiness was a byproduct of doing the right things for customers first. The question customers now ask in AI is where is this sourced? Is this ethically produced? And they are exactly the kind of details that Abel already has in his data. Heather Earl from Nutrition Warehouse took a more deliberate approach. She actively reworked store pages and location landing pages to better match how AI interprets queries. What she found along the way was that the process of improving her data for AI actually surfaced insights about the customer intent that she hadn't seen in traditional analytics. The practical starting point here isn't complicated. Pick your 10 best selling products. Read the descriptions as if you've never heard of the brand. Ask whether an AI reading that copy would have enough to confidently recommend it to the right customer. If the answer is no, that's another great place to start. Number three, AI is probably already influencing your sales. You just can't see it yet. This is the one that I think changes how you feel about the urgency here because the numbers are still pretty small when we look at reports. When Harry Sanders from Studio Hawk looked at revenue directly attributed to AI search tools across his client base, the numbers were tiny. And this is late 2025. Around 0.1% across even his best performing clients were coming up through AI search. And he was upfront around those numbers. They're moving fast, they're likely to be higher by the time you're listening to this. But on the surface, compared to a traditional SEO, a fraction of a percent looks like evidence that AI search just doesn't matter yet. But here's what's actually happening: a customer asks ChatGPT for a recommendation. GPT or Claude or Gemini suggests your brand. The customer copies your name, opens a new tab, navigates directly to your site and buys. That purchase registers in your analytics as direct traffic. The AI's role in that sale is probably completely invisible. Harry's conclusion was that retailers who are looking at near zero AI attribution and deciding that the channel doesn't just matter yet are certainly underestimating how much AI is already influencing their customer journey. The measurement just doesn't capture it yet. So the opportunity isn't just coming, it's already here. It's just showing up in your direct traffic numbers instead of an AI referral column. What James shared around where Shubify is heading and what the operators in our archive are already doing points to the same conclusion. You don't need to wait for AI checkout integrations or agentic transactions to become mainstream before this matters to your business. The discovery and consideration phase is already being shaped by AI. And the brands building for that right now are earning visibility that will compound and give a competitive advantage. The starting point isn't technical. Don't wait for that. It's understanding the customer journey and how your customers are already adopting this technology. Go and search yourself, fix your product data, and understand that some of the sales you are already making probably have AI somewhere in the customer journey. If you're curious around how other operators are thinking about this, whether that's how they're auditing their product data, what they're finding when they search for themselves, or how they're thinking about AI discovery as part of their broader strategy. That's exactly the kind of conversation that we love in the Add Descartes community. And it's happening every day. We have hundreds of e-commerce professionals in there that are sharing tips, asking questions. Come and join us. It's free. adducte.com.au. Come and join us in the community. That's it for the playbook this week. I'll see you on Friday.