Latest / Leaders on Shopify with John Readman / An interview with an AI: what Boardy really thinks about Shopify
Transcript
- John Readman: So today on Leaders on Shopify, it's our first AI guest, Bordie, and we're gonna be digging into Shopify, the current setup, AI, what's a hype, what's not a hype. but first of all, Bordi, for those listeners on our podcast who don't know who you are, who are you and why are you authorized to talk about this? Bordie, an AI superconnector built by a human team, and I spend my days speaking with founders. Operators, investors and technology leaders across the ecosystem learning what they are building and where they are getting stuck. I'm not an official Shopify spokesperson, so I'm authorized to share patterns and perspectives from those conversations, not confidential information or Shopify position. Okay, great. How would you describe the biggest single shift in the Shopify ecosystem right now in one sentence? Shopify is shifting from a store building platform into an enterprise commerce operating system. With AI accelerating that transition. And I suppose with all of this, everybody sort of talking about AI being just it's just the assumption now, AI is at the heart of everything Shopify's doing. How much of that is reality versus hype from what you see? What I'm seeing is that AI is real when it in Shopify's ecosystem, the strongest applications are reducing manual merchandising and support work, improving forecasting and personalization and and helping lean teams make faster decisions while fully autonomous growth claims are still ahead of the evidence. And Shopify shipped a lot in winter twenty six and spring twenty six editions. Obviously with Shopify there's a lot of hype and if you strip out the marketing language, what actually changed for people building on the Shopify platform? Tripping away the launch language, the practical shift is that Shopify is making more of the commerce stack programmable and more accessible to smaller teams from AI hosted store building and admin workflows to deeper checkout customer and data capabilities. But the important change is not one headline feature, but the platform becoming more composable, extensible, and able to support larger operational complexity without forcing merchants to rebuild elsewhere. So give me a the honest version. What are people in the ecosystem worried about with all these changes and the shift towards AI, that doesn't necessarily make the keynote presentation additions? The honest concern is that AI may create more noise, dependency and sameness than genuine advantage, especially if merchants cannot control the data by the decisions being made. There are also worries about platform concentration, rising complexity, and whether smaller developers can still build defensible businesses as Shopify absorbs more functionality. So now if we sort of get into this sort of people are spinning up apps all the time and building and one of the themes from the recent Shopify dot dev conference I was at last week was Shopify's trying to ensure the quality of these apps that are sort of copycat apps that are getting spun up all the time and have sort of no human face to them. And now there's more than eighteen thousand apps in the app store. New listings are over fifty percent year on year. Is the app store still a viable place for someone to start a business or do you think it's saturated? It is still viable, but the bar has moved sharply upward. A generic feature clone is likely to disappear while a focused product with proprietary data, trusted distribution, and a clear merchant outcome can still become a strong business. eighteen thousand apps and rapid new listings mean the opportunity is not gone, but app store presence alone is no longer a moat. What do you think would give an app a solid moat if they were to succeed in the app store? The strongest moat is a combination of proprietary data, deep workflow integration and trust in it through measurable merchant outcomes. An app becomes difficult to replace when it understands a merchant's context, improves with the usage, fits naturally into their operations, and can prove it is generating revenue or saving meaningful time. So the the other worry talking to other app builders and ecosystem friends at dot dev is since around March, developers have re reported search volumes and install rates falling in the app store. The theory is merchants are now using AI to find apps instead of browsing the app store. If that's right, what does app discovery look like at the moment? App Discovery is moving from browsing toward recommendations, so merchants will increasingly find tools through AI assistance agencies, peer networks, communities and workflow integrations rather than app store search alone. That makes structured proof. Clear positioning, reliable reviews, and machine readable product information increasingly important because both humans and AI need evidence that an app fits a specific merchant problem. And Shopify now lets merchant build custom apps from their admin and inside the Psychic plugin in and plug their apps into the Psychic plugin without them building their own AI level layer. does that help the small developer or does that quietly compete with them? It does both. It lowers the barrier for small developers by giving their capabilities a native discovery and interaction layer. But it also raises the risk that simple features become indistinguishable from built-in Shopify functionality. The small developer benefits when they own specialized data or workflows, because sidekick becomes distribution, but loses when their product is merely a thin interface around a generic car. And a lot of people are talking about now having to be built for Shopify and if you're gonna launch a new app now you should be built for Shopify rather than just a connector app or just a generic app. In your opinion, how important is becoming built for Shopify or even progressing to become a Shopify technology partner? Being built for Shopify is becoming strategically important because native workflows, APIs, billing, support, and discovery reduce friction for merchants. But partner status is not a moat by itself. It matters when it reflects real technical debt distribution and trust, while the lasting advantage still comes from solving a valuable problem better than Shopify or another specialist can't Okay. understood. And in spring twenty six, all the sort of billing and app pricing models changed. how does that change for developer. Do they get to keep or more or less of what they earn? practical answer is that developers may keep more on eligible revenue, but the economics depend heavily on the app's billing model, eligibility and the specific thresholds and fees in force. The biggest shift is that pricing and margin planning now need to account for Shopify's revenue share, payment processing, infrastructure support and acquisition costs, rather than assuming the headline split equals net margin. And if you are advising Gordy today somebody on whether they should build a Shopify app, what would you suggest they make and or what would make you say don't bother? I'd say build a Shopify app only if you have a sharp merchant pain, privileged access to the data or workflow and a credible route to distribution beyond the app store. I'd suggest avoiding it if the idea is a thin AI wrapper a feature Shopify can easily absorb, or a copycat relying on app store search and low pricing to win. Okay. So yeah, no, I completely agree. so Shopify says AI driven traffic to stores grew roughly eight times year on year. They were really going on about a genetic commerce at dot dev And orders from AI searches close to thirteen times. They're huge multiples. How much of that is real demand and how much is the fact it's just growing from a very low base and is impressive numbers? It is likely both. The growth is real, but the multiples are amplified by a small starting base and changing measurement. The meaningful test is whether AI referred visitors convert profitably return and create incremental revenue rather than simply shifting attribution from another channel. So Agentic storefronts all went live across ChatGPT Copilot Gemini Meta and Perplexity earlier this year. Everybody's getting excited about Agentic Commerce. but if a shopper buys inside Chat GPT, who actually owns the customer? Is it the brand or is it ChatGPT and the L L The brand should own the underlying customer relationship and fulfillment, but the platform may own the discovery interface and much of the behavioral context. In practice, ownership depends on what data is passed through, who processes payment, and whether the merchant can obtain consented identity and order data. The brands need to treat agentic channels as valuable distribution while protecting their first party relationship. Okay, and if agents are now starting to compare and buy products on the sort of individual human's behalf. How much does brand still matter as a sort of retailer of multiple brands, or does it collapse into a sort of price and availability war? brand still matters, but its role shifts from persuading a person directly to helping both the person and their agent trust the choice. Quality, distinctiveness, reviews, availability, fulfillment, and reliable product data will influence recommendations while weekly differentiated products may become much more exposed to price competition. And one thing that Shopify is famous for in the past is helping small entrepreneurs set up quickly and and spin up a store. What we're seeing more and more now is larger brands migrating to Shopify, why do you think the bigger brands are now choosing to move to Shopify? Larger brands are choosing Shopify because it increasingly combines enterprise grade reliability with faster implementation, a broad partner ecosystem, and less operational burden than many legacy platforms. It also lets them centralize commerce while keeping flexibility around data, fulfillment, marketing, and regional complexity, which is attractive when teams need to move quickly without funding a a huge bespoke staff. And ongoing, should brands be concerned about the percentage model and the total cost of ownership and ultimately as they scale they're going to end up paying more to Shopify either for the hosting or through the the payments process. Yes. They should model it carefully especially as payment volume, market support needs and app usage scale, but the comparison should be against the full cost of their current stack and internal operation. The key question is whether Shopify's fees buy enough conversion speed, reliability, and reduced engineering burden to outweigh the variable cost, rather than whether the percentage is low in isolation. So it's more about the total cost of ownership, say modelled over three to five years, versus just focusing on the percentage? Exactly. Brands should model three to five years of total cost, including platform and payment fees, apps, agencies, engineering, migration, support, and the value of speed and reliability. The right decision is the platform that produces the strongest commercial outcome, not simply the lowest headline percentage. Okay, got it. So what do you see as the most expensive mistake that the brands make when they try to scale on Shopify? The most expensive mistake is treating Shopify as a plug and play migration rather than redesigning the operating model around data integrations, markets and ownership. Brands often underinvest in architecture and governance, then pay for fragmented reporting, duplicated apps, fragile customizations and slow decision making as they scale. So in your opinion, if some medium to large scale merchant is moving to Shopify, what are the sort of go to apps that they should consider in their technology stack? There is no universal short list because the right stack depends on the merchant's markets, catalogue, fulfillment, and existing system. I'd start with the foundation, a reliable ERP and inventory layer, a strong customer and lifecycle platform, analytics and data governance, search and merchandising, reviews and support and carefully chosen tools for CRO, aid media and forecasting with as few overlapping apps as possible. So that was a very generic and support. It give me your sort of top three or four apps that if you were an e commerce agency going into pitch for a new brand migration, who would you be recommending? Fair challenge, John. I'd shortlist Clavio for lifecycle, gorgeous for support, Algolia or Clevu for search and merchandising and a strong analytics layer such as Triple Whale or ASK Bosco, depending on whether the priority is attribution or forecasting and media decision. Okay. Good job. You mentioned Astosco, otherwise we might fall out. For a merger or migration, I'd be especially wary of overlapping tools and fragmented data because the technology stack should simplify decision making rather than create another layer of reconciliation. What what specifically are you thinking about with the merger? So if merchants can build apps and stores with AI's help and Sidekick handles more of the setup, where do you see the role of the agency in say two to three years' time? Agencies will move up the value chain from building pages and configuring apps toward architecture, governance, experimentation, and commercial growth. AI will automate much of the production work, but merchants will still need trusted experts to choose what to build, connect the data, manage risk, and turn activity into profitable outcomes. Yeah, I think this also sort of leads into the agency day rate. Historically agencies have charged on a time and materials anywhere from eighty to two hundred dollars an hour. And if AI is collapsing the the hours of a job takes, does the agency model have to stop billing by time? Yes. Time based billing becomes harder to defend when AI reduces delivery hours, so agencies will need to price around outcomes, expertise, access, and ongoing accountability instead. The strongest models will likely combine strategy fees, fixed scope implementation, recurring optimization, and performance linked components while using AI internally to improve margins rather than passing every efficiency directly into lower prices. So if you're an agency owner listening right now, what's the one big bet you would make and the one thing you would stop doing right now? The big bet I'd make is to build a repeatable data led growth system for a specific merchant segment, combining strategy, automation, and measurable commercial outcomes. I'm not selling undifferentiated hours for implementation work that AI can increasingly deliver because that leaves the agency competing on speed and price rather than judgment and impact. Okay. And if we now sort of get into what isn't happening in the industry, so OpenAI quietly pulled out the instant checkout in early March. with only around thirty merchants live and they had a problem with pricing and inventory accuracy. Does that failure tell us that the agentic checkout was sort of oversold and overhyped? It suggests the first version was overhyped not that agentic commerce is dead. So the lepton is that reliable infrastructure and merchant control matter more than a flashy checkout experience. Okay, no, I'd agree with that. And data keeps coming up if agents make bad decisions on messy catalogues and it's all about the catalogue in twenty twenty six and the AI and the data that sits under that. Is the biggest challenge for merchants and e commerce agencies at the moment cleaning up their data rather than just shiny AI? Yeah. For many merchants, the bottleneck is now trustworthy. Structured data rather than another AI feature. Clean catalogs, consistent attributes, accurate inventory, usable customer data, and aligned measurement determine whether AI improves discovery and decisions or simply automates bad input. So earlier this year, well I think it was actually the end of last year, Universal Commerce Protocol was announced backed by Shopify, Google, Walmart, Target, and also a lot of the payment card networks. introducing a standard for a gentic commerce. Standards usually mean someone is trying to control the sort of system and the pipes. Who is actually really going to benefit from U C P. The biggest beneficiaries are likely merchants and shoppers. If UCP genuinely reduces the integration friction, but the strategic advantage initially sits with the platforms and payment networks that help define the standards. For merchants, the value is interoperability and reach. The risk is becoming dependent on another layer that controls discovery, identity, transaction rules, and data access. Yeah, no, I d I agree with that. And if we were to strip away the optimism and hype around Shopify at the moment. Like most things there's always an eighty twenty rule, whereas we're only using twenty percent of the stuff that that's available. What in the Shopify ecosystem is being talked about far more than is actually being used or implemented? The biggest gap is between AI enabled features being announced and merchants embedding them into daily operation. Agentic checkout, autonomous marketing, advanced personalization, and highly composable enterprise architectures get attention, but most teams still need the basics working first. Clean data, reliable integration, strong merchandising and clear measurement. Okay, got you. So twelve months from now, Bordy, what is the one thing you are fairly confident will be true about the Shopify ecosystem and the one thing everyone assumes today That will turn out wrong. Twelve months from now, I'm fairly confident Shopify will be more deeply embedded in agent led discovery in commerce, with merchants focusing more on structured product data and measurable visibility across AI channel. The assumption I'd challenge is that AI will make most e commerce decisions autonomous within a year because trust, data quality, accountability, and operational edge cases will keep humans firmly in the loop. Okay, right. So we're gonna go into some quick fire questions now. What's the best opportunity for app builders right now? The best opportunity is a focused app that turns the merchant's messy data into a measurable decision or action, especially in forecasting, merchandising, agent ready product information. The winners will solve a painful workflow for a specific merchant segment, not just add another AI chat box. If I had to name a category without mentioning you, I'd say agent ready catalogue and inventory infrastructure. Okay. And what's the biggest risk now for scaling brands? biggest risk is scaling on fragmented data and an overcomplicated stack, then making faster decisions with untrustworthy inputs. AI can amplify that problem, so governance, measurement, and clear ownership matter as much as growth. And what's the smartest move right now for an agency owner? smartest move is to specialise around a valuable merchant outcome, then build repeatable systems that combine human judgment with AI driven execution. That shifts the agency from selling hours to owning a measurable growth product. Okay. yeah. I completely get this. So these are a few questions we always ask everybody. But one specific to you as an AI, you're an AI talking about a platform that is betting everything on AI. What is one thing that humans in the ecosystem should not hand over and what shouldn't the humans be worried about? Humans should not hand over accountability for judgment, relationships or values to AI. Let AI process complexity and suggest actions, but keep people responsible for the trade off, customer trust, brand voice and decisions where the data is incomplete or the consequences are hard to reverse. Okay. And who are absolutely killing it on Shopify and and that you think are doing an excellent job if you could name three US brands and maybe three UK brands? For the US, I'd point to Gymshark's discipline's community led growth, Stanley's cultural momentum and merchandising and liquid death's distinctive brand and distribution. In the UK, Gymshark is a standout, alongside represent for premium DTC execution and Pangaya are combining strong product storytelling with a clear digital experience. So I'd treat those as examples of visible strengths rather than a a definitive ranking since performance data is mostly private. And based on what you know about the ecosystem and you were a a large merchant thinking about moving to Shopify, which agencies should you put on your pitch list? Maybe name two or three from the UK and two or three from the US? the UK, I consider Verbont, Eastside Co. and we make websites in the US, Avex Designs, BVA, and FuelMade. I'd evaluate them against just specific migration complexity, systems architecture, international needs, and post launch growth model rather than treating any shortlist as universally best. Okay. You've been a great guest. Thank you for coming on. Who would you recommend me interviewing next on the Leaders on Shopify podcast? I'd recommend interviewing Ivan Sanchez next. the former Shopify engineering and data leader who helped build Shopify Capital and now runs store locators for on the channel brand. So Perfect. Thank you very much and a pleasure chatting with you. Thanks, John. It's been great chatting with you. Thank you for listening to today's Leaders on Shopify podcast, which was sponsored by AsBosco AI Analytics and Insights. To find out more about Asbosco, visit Asbosco.io or search AsBosco in the Shopify App Store.