Elevate — The Global Commerce Summit 2025 | Dirk Hoerig, commercetools

Clips
More from Elevate — The Global Commerce Summit 2025

Dirk Hoerig

Founder

commercetools

play_circle_outline The Shift to Digital Payments: Driving Revenue and Optimizing Margins in a Unified Commerce World

play_circle_outline Growth of agentic shopping driven by AI assistants taking on more autonomous tasks.

play_circle_outline Challenges for brands and retailers in adapting to agentic shopping and building brand engagement.

play_circle_outline Introduction of Model Context Protocol (MCP) for standardized access to retailer systems for agents.

play_circle_outline Benefits of MCP in enabling collaboration, sharing context, and facilitating agent-to-agent interactions.

Info
Transcript

Dirk Hoerig, commercetools

Dirk Hoerig

Founder commercetools

Dirk Hoerig, founder and chief information officer of commercetools GmbH, joins theCUBE’s Scott Hebner and Jennifer Wu at Elevate — The Global Commerce Summit, for a timely discussion on the future of digital commerce. From unified architecture to the rise of intelligent agents, the session offers an incisive look at what’s reshaping consumer expectations.

The conversation touches on topics such as digital payments, the evolution of unified commerce, and the projected surge in agentic shopping. Hoerig unpacks how these shifts are driving a fundamenta... Read more

explore Keep Exploring

What are some trends in digital payments and how are they impacting the world of unified commerce? add

What is the current estimated global digital commerce GMV and what is the projected growth rate for the next five years? add

What are the important steps to take in order to successfully grow a business in a market that is becoming more open and requires new thinking? add

What is the purpose of MCP, the model context protocol, in relation to providing access to a retailer system for agents? add

What does the MCP allow different specialized commerce agents to do? add

bolt Powered by CUBE AI

Dirk Hoerig, commercetools

search

Patricia Wu

>> Welcome to ELEVATE, the Global Commerce Summit. We're here at the Miami Beach Convention Center. I'm Patricia Wu, your host, along with my co-host and analyst, Scott Hebner. So Scott, are you ready to dive into the crystal ball of commerce?

Scott Hebner

>> I am. We've had many fascinating conversations so far. So now we're going to go a little deeper, I think, into technology.

Patricia Wu

>> Oh, yes, we are. And we're looking more into the future. And we've got the visionary founder and chief innovation officer, Dirk Hoerig from Commercetools. Thank you so much for joining us.

Dirk Hoerig

>> Thanks for having me with you.

Patricia Wu

>> So let's just dive into the future right away. What are some key trends that you're looking at?

Dirk Hoerig

>> Yeah, we obviously see a lot of things changing in the market at the moment, and we will talk about the elephant in the room, AI, I think in a couple of minutes. But before we get there, let's talk about a few other trends that we're actually seeing. So first, digital payments. You could argue digital payments in itself, especially for e-commerce, it's a natural thing. How should you pay online otherwise if it's not digital? But it has evolved significantly over the last couple of years. And in the world of unified commerce, because we are not thinking in silos at Commercetools, and with our customers, we are thinking across all of the touch points that actually their consumers are interacting with the brands and retailers that are running on our platform. Their payments is a critical element of identifying the customers everywhere on helping driving top line revenue. But on the other end, also for the business of brands and retailers to optimize on gross margin. Because if you have the wrong payment setup, you can also lose a lot of money. And if we talk now about some trends where we see cash significantly disappearing. I just yesterday looked into some recent e-marketer stats on various payment methods. Over the next three years, cash payments all over the world, the number is actually stagnating. And in ratio to all of the other digital payments like credit cards, wallets, and so on, they are, you could say, even shrinking because they're growing significantly in comparison. And you could see that as examples by some mega events. Tokyo Expo this year, no cash, all digital. Paris last year, Olympics. I experienced it myself, all of the venues, all of the payments, no matter if it was for drinks, for food, if it was for merchandising on each of the sports places, all just digital. And not just for the online touch points, but actually for all of the little booths that had been around all of these venues. So there's a significant increase in demand and change out there, and an opportunity for customers to grow top line business by having the right payment methods, having the checkout flow easy across all touch points, so both online and offline. And that's definitely one of the trends that we see coming up.

Scott Hebner

>> Yeah, the loyalty programs are driving that too, right?

Dirk Hoerig

>> Absolutely.

Scott Hebner

>> It's like hate when I have to pay. There's a couple bills I still have to write a check for, and I'm like, "I'm losing points."

Dirk Hoerig

>> Yeah. And here's the thing, right? Loyalty program, many people think standing in line or standing in front of a checkout. And the thing that you don't want to do is you don't want to register with your email address again. You stand there, then they're you're asking you, and then you start to spelling it. And so it's just a little bit taking time. It also has some risk of typing some errors in then you're not getting the points so it's not connected. And the easiest way to really get the loyalty programs across all touchpoints together and have it in a convenient fashion for you as a customer is actually identifying through payments. So therefore, you need to have a holistic payment solution that actually can combine all of these touch points.

Scott Hebner

>> Interesting.

Patricia Wu

>> Yes. How many times have I stood there? The phone number and the email with the little stylus? But there's something that you said in your keynote this morning about agentic shopping, and it would be 30% of online commerce by 2030. That seems like a mind-boggling number, but you were saying it may be low.

Dirk Hoerig

>> It actually is. But follow me here. Today, '25, we have about 1 billion monthly active users across Gemini, ChatGPT, Perplexity, and some of the other AI assistants. So that's the starting point, 1 billion users globally. So roughly, what is it? 11-12%, 13 of the population already out there using them on a general basis and got used to these technologies. So this goes far beyond the hype or a Silicon Valley trend. We said, "Okay, yeah, technology-driven people only have it." We're beyond that and the adoption curve. Second, when we looked at growth and adoptions of new market-changing technologies like mobile internet. So we saw with the rise of the iPhone, smartphone then 3G, 4G, 5G, mobile internet in the US in 2010 was somewhere around 2%. 98% of the internet traffic had been consumed by desktop devices. Six years later, 2016 we'd been at 50%. So mobile caught completely up with the desktop and had been on par. That took six years. But we didn't have 1 billion devices out there in 2010, it was significantly less. It had been a couple of million that had been distributed and used within the market. So now the starting point is 1 billion. Everybody's rolling out shopping capabilities within their assistants. We yesterday saw from our partner Google great announcements at their IO events on the shopping side, which will drive customer behavior over there. A little bit less from browsing on the site versus directly interacting with an agent. And now if we think back about this fast growth of the last couple of years on trending technologies, this one here is growing even faster. So therefore, the prediction is we're now at roughly $6.8 trillion commerce GMV, digital commerce GMV globally. And yes, agentic shopping is just starting. But if we apply only the past numbers, and I think there are a lot of statistics out there that tell us this one here is growing faster, but only if we apply the last trends of internet, of mobile commerce and some of the other things that fundamentally changed our market to what's now happening, then five years from now on getting to 30%. Which it's a big number because 30% might sound, "Yeah, okay, it's already a bit," that's $3 trillion US dollar of GMV. That's all of e-commerce worldwide in the year 2018. And that's where we are believing that this drives significant impact for all of us, both on the consumer side, but especially and even more if you're a brand or retailer and need to figure out, hey, if my best buyer is an agent, how do I deal with that?

Scott Hebner

>> And so a billion people are using AI assistants. And a lot of them using it obviously, to shop. So what is an agentic shopper? How does that take it to the next level? Basically, how do you define the difference between an AI assistant and an AI agent in the context of commerce?

Dirk Hoerig

>> Yeah, that's a good question. So the term, agentic in itself is so new, I don't know how it is for you, but when I'm typing it into my computer, it always is becoming shown as a spelling error and I need to edit it.

Scott Hebner

>> Magentic. Yeah.

Dirk Hoerig

>> Yes, yes. It always wants to change it because it's so new. And there are definitely some definitions out there, I think for people to understand it. It's somebody that is acting autonomously on your behalf. So guided where you as a human, let's say give it advice, "Hey, I'm interested in a specific product," or "Please look this up for me." And then you can give it even a task plan and a, "Hey, don't buy that yet, but please get back to me once you find it at a specific vendor, at a specific price point with all of my preferences and come back." And then the agentic, which is acting on behalf of the human, will actually start processing it. See it as your personal assistant that's always nearby and helping you making the right buying decisions. And the great thing there is it knows much more about your preferences than any brand or retailer in itself could know because they are limited within the knowledge that they have about you when you're browsing a site.

Scott Hebner

>> The way we've pieced it together is that an AI assistant is basically task-based. It's going help you execute a task, particularly those that are easily repeatable. And as you said, an agent gets you into helping you achieve a goal. I want to buy something. And it goes through off autonomously and helps you achieve that goal, creates a plan, so on and so forth. And then agentic AI is when you start to wire agents together to help solve more complex problems or to work as a workflow or to take advantage of different agents who have different skills. So how do they work together to solve a problem? Is that-

Dirk Hoerig

>> Yeah, I think that's pretty-...

Scott Hebner

>> what we're going to see here?...

Dirk Hoerig

>> good description.

Scott Hebner

>> Yeah.

Patricia Wu

>> Well, what I heard was I may have help coming my way for my questionable fashion choices.

Scott Hebner

>> Yes.

Dirk Hoerig

>> Well, systems-

Scott Hebner

>> It's going to tell you the truth.

Dirk Hoerig

>> Yeah, it will definitely give you an opinion, but it will also make things easier. And as we have shown here earlier today, at a pretty good example, it also helps you comparing products significantly easier when you struggle on making a choice. And fashion is maybe pretty much based on your personal preference and not too much just on technical criteria. But when you're thinking about just you're going on vacation and you want to buy a new camera, and so the options that you have out there, it's overwhelming. And yes, you can do faceting search, you can try to compare all of these criteria. But with an agentic shopper, you basically get the expert of these products in front of you, which can help you drive the right decision for the use case that you're actually having.

Patricia Wu

>> Which really helps the brands and the retailers because that agent knows you better than they do. But you said something earlier. Like what do you do if your biggest buyer is an agent? How do you deal with that?

Dirk Hoerig

>> Which is actually a challenge that over the next few years, almost all of the brands have to face because we need to differentiate between two different purchasing flows. The one that is within brand. And within brand means you are visiting their website or their app and they have an agent running, which is important. Most don't have that yet. Or we know the little chat windows, but it's not about the chat window where you have a little support floor running. It's really about having somebody who can create a personalized experience in real time. That's in-brand agentic shopping. They will launch and provide you with an agent that's actually helping you procuring within their product offering catalog, knowing obviously your preferences to the degree that they have stored it in your customer profile. When we talk about agentic commerce these days, we often mean the in-chat flow on the external assistants, so within a Gemini where you're not visiting the brand. And this is also what we meant when we spoke earlier about the 30%. So not happening on your site. Where you're never visiting the website, where you're never visiting the app, but you are interacting with indirectly through the agent with the brand. And the challenge for them will be is how do I do brand building? How do I actually engage with the customer when I don't have the experience under my control anymore because it's not happening within my territory? Think about it like you have your store, but it's not somebody who's coming into your store and buying it. They buy with somebody who they are meeting outside of the store somewhere in a park on the street and just tell them. So maybe your assistant and say, "Hey, okay," I forgot my ... Actually, a real use case. I didn't get my luggage in time here for the event. It came on actually on the minute, but I was thinking about maybe I need a suit on time because just with the clothes that I came here to Miami, that wouldn't work out well. So I would have, because being limited on time, sent somebody on buying something for me. And now I think about this human driven approach, fully AI driven. And this is something that everybody now needs to figure out. So to your question, how do I start here? Is building the fundamentals that you first can participate in there. Because if you are last to come to the party, you might not find the right partners for the dance anymore. Might be a stupid kind of scenario or example here to take, but it is pretty much like that. With a locked in strategy, you will not be able to successfully grow your business in now a market that becomes significantly more open and where you need to actually embrace that new thinking. So that means partially investments in technology, but that's not the big kind of thing on opening up. We launched actually today our Commerce MCP. I think we'll talk about that in a minute, what this means and why this is helpful. But actually opening up to this agentic environment and help driving for the customers, the use cases that actually create business value for them. So don't do agentic because it's just a hype topic. Let's not repeat the digital transformation trap that many companies said, "Oh, we now have to do this. Let's take our old fax-driven processes and transform them to email or something else." That's not about it. Think about different. How can I drive customer value differently because everything is going to change? This is not a calculator or Excel. And before that, we did it in a different kind of way. We are now fundamentally changing or we'll see changes over the next couple of years how we inform ourselves, how we are being creative, but also how do we buy things.

Scott Hebner

>> So let's go to some of the news. A big part of the news, which was as you mentioned was MCP. So the way I put all this together from watching the keynotes, and first of all, architecture matters as we go forward given the rapid innovation cycles. And we were talking to Andrew earlier, you have mock as a foundation, so it's microservices, it's a API based, it's cloud native and it's headless. So you have that architecture that allows you to move rapidly and infuse new technologies. On the very top, you have an ecosystem of partners and contributors, which over time will be building more and more agents. You announced MCP, the model context protocol. How does that fit into the whole picture here?

Dirk Hoerig

>> It is a standardized way to give agents out there an access to a retailer system without having for every agent to manually do that again. So as you said, the fundamentals that you need to have in place is an API. If you are not having an API, if you're having a closed system and you're integrating some agents there, this will not scale up, especially not in a partner ecosystem. The model context protocol had been developed so that not every company is doing their own kind of thing. And I think a good example here is USB. We've seen about, I think 10 different formats of USB interfaces in the last 10 years. MCP is a standardized way that we are not seeing oh, vendor A is doing this versus vendor B is doing that and then the agents out there are starting to struggle to adopt the information that they are retrieving. So giving them a standard way that as a brand or retailer, you don't need to do custom development. So that for everybody who's providing an agent on your solution, it gives them an easy access. And on top of that, Commerce MCP that we launched, we're having our API hub. And that's actually where all of these agents are being listed, where they're living and where our customers can choose from. And that can be those that help you on leveraging shopping capabilities, but it can also be helping you doing price optimization, catalog assortment, optimizing processes within your backend, analyzing for good or bad promotions that you're running. So it will be an agentic ecosystem that is open for our partners and for our customers to develop that they can actually pick and choose from.

Scott Hebner

>> And the MCP allows these different specialized commerce agents to actually interact, collaborate, share context so they can make decisions-

Patricia Wu

>> The same language.

Scott Hebner

>> Yeah, right?

Dirk Hoerig

>> Yes, absolutely.

Scott Hebner

>> So that adds to the architecture.

Dirk Hoerig

>> And then we have agent to agent. So this is also something that is going to happen, that it's not only just one agent then interacting on the one end, here's the brand and there's the consumer. But the brand is providing an agent and the consumer is having an agent or even probably another retailer in a B2B scenario. And then the agents in itself are dealing with the other. And the MCP is actually helping all of these that you have a standardized way and that you don't need to do single development agent by agent, AI model by AI model. And the MCP also allows you to switch the AI models down below on the runtime. Because that also has been a big question for many companies out there. If I now invest into AI, in I don't know, can be OpenAI, can be Gemini, can be something else, every couple of months there's a new model out, which is deprecating the value of the old model. But you started to train on top of that, build with that, work with that, and then it's, "Oh, it's all of my investment being gone." The MCP allows you to switch the models down below out easily, so you don't have any model dependency on the own agents that you are building because it's a standardized way how to interact with it.

Scott Hebner

>> Enhances the composability element, right?

Dirk Hoerig

>> Right.

Scott Hebner

>> Do you have any customers or ecosystem partners that are starting to use MCP?

Dirk Hoerig

>> Yeah. It's early on, but we see a great traction across our customers. We actually have had PetSmart also here at Elevate. Greg, their CTO, gave a good example about their newly launched loyalty program as we spoke about that earlier, because there's also significant AI value in there on driving customer recommendations, personalization and so on. That's a good AI example of a customer of ours.

Patricia Wu

>> So it sounds like the future is adaptability, flexibility, connection, collaboration from digital payments not being siloed to agents.

Dirk Hoerig

>> Absolutely. And the term that I always like to use here is unified commerce, because unified commerce is not only combining online and offline. So that's also, it's all of the customer touch points, including the new customers, the AI customers that are coming on top of it and harmonizing it on one infrastructure layer. I think that is kind of the goal. Because if you're not having that, as you said, you run into silos and you will have broken processes, it will run inefficiencies. It will not lead to great customer experiences, and that's what we're looking at.

Scott Hebner

>> It's powerful because it's unified, but it's also future-proof.

Dirk Hoerig

>> Yes.

Scott Hebner

>> And you can adapt quickly. Like you said, you have one agent that maybe is no longer given competitive advantage, but there's another agent that you can plug in that does it.

Dirk Hoerig

>> Absolutely. And who knows how the next couple of years will look like? Just if we look at the last five years, everything that happened and what changed. We always thought, okay, there'd been COVID things changed. We all as humans adapted. And then, okay, the last two years had been crazy. Then the next thing is coming, then the next thing is coming. So pace of change just continues to accelerate. And as you said, having a future-proven setup, future-proven technology that gives you this unified ability to harmonize all these channels, even if there's more coming or especially if there's more coming, that's the key to success, to be actually ready and prepared and benefit from what's going to come.

Patricia Wu

>> Yes. Thank you so much, Dirk, for giving us these tools and insights into navigating this rapidly evolving world.

Dirk Hoerig

>> Thank you so much for having me here.

Scott Hebner

>> Fascinating.

Patricia Wu

>> Yeah, thank you. I'm Patricia Wu, you've been watching ... And Scott Hebner. How could I forget my co-host and analyst-

Scott Hebner

>> Right here....

Patricia Wu

>> right here? So you are watching our coverage of ELEVATE the Global Commerce Summit. We're going to have a lot more coming your way. And you're watching theCUBE, the leader in enterprise tech news and analysis.