Business Growth with AI Innovation

Transcript

Hi, this is Russell with Adobe Commerce. This is the first video out of three of a previously recorded webinar. For now, we’re just going to continue with this session on business growth with Adobe AI Innovation. The biggest shift that we are making in Adobe Commerce team and in Adobe Commerce engineering is prioritizing that you be successful with the SaaS platform. First, that you can come on the SaaS platform easily, make your migration easy. There is a full section on that. And the second one is those of you who are taking the leap.

Almost 40% of our engineering is transforming itself to become forward deployment engineers, which means it is not Adobe support which is going to help you. It is our ultimate engineering team that will engage and ensure your goal lives happen and happen in time and whatever comes in the way, we are committed to fixing it. And a lot of things like, you know, making reporting even better, customer facing observability, supporting a unique product type, search and recommendations functioning a given way. All these are some of the examples that would be done as capabilities in the product, but to ensure that your goal life goes as planned.

All right, without much delay, let me now take you into the world of what does our AI roadmap look like. So first of all, things have changed and things are changing for us very fast. We are almost in a phase where the shifts in the tech industry, we can call them tectonic and e-commerce is no exception. We expect at Adobe, we expect that a significant part of traffic, e-commerce traffic is now going to be agentic. What is that percentage going to be and how soon that change is going to happen? We don’t know, but we do know that that future is coming. It may come sooner, it may come a little, it may take a little longer. How do we say that? Our own research, and this is done at Adobe level, says that Gen AI traffic, which is me and you using LLMs to retail sites, traffic driven through because of the work we did or research we did in LLMs has surged 4700% year over year. AI referred shoppers such as you and me spend 32% more time on sites. We view 10% more pages, we bounce 27% less. So for all our customers, they are looking for a solution which is optimized not just for humans. In today’s world, they are looking for a solution that is optimized for AI agents as well. Many of us were part of the whole digital transformation wave that happened. We are now part of this transformation of agentic transformation. The transformation is happening on three big fronts in the context of e-commerce. First, the discovery itself has changed. We are not discovering as much with a social engine or using social media. We are discovering more using LLMs. The browsing experience has changed because it’s not just humans that browse, agents are browsing. And the biggest change in my eyes is the fact that the experience itself has changed because earlier we used to go on a search engine which used to give us results and redirect us to someplace else. Now when we are in a LLM interface, we don’t go anywhere else. We just stay there till we get what we want. So the entire experience from the beginning to the end is now in one interface. And that is a key point I want all of us to keep in mind as we go through this 45 minutes together that what does this change in paradigm mean that any journey is beginning and is going to end in that same interface. And how does this impact the whole e-commerce offerings that vendors such as Adobe and others will make and how will it change the whole experience that our customers will need to offer. So what are customers telling us? First thing everybody is telling us how can we make our product show up in LLMs better. Because what do LLMs do currently is quite opaque. So let’s say if I am a handbag company, Gucci’s handbag and Chanda’s handbag, they are on the same level playing field. So everybody now has the problem how do I get my product to show up in the right way. Now LLMs are becoming a primary channel for many of you, right, for many of our customers. So how best to actually function in this new expanding channel? How do we expand to new markets in this new paradigm? How can catalogs, storefront, you know, all those be taken to a new market much faster. Let’s say you are gap.com a much higher percentage of visitors coming to your storefront are now much informed they have done their research. So as gap what do you want to do? You want to give them what they want really quickly. You want to have a data driven approach where they know where you know what they want and you can serve them very quickly. And then how you can leverage that to sell them a little bit more or a little bit related items.

If you are a B2B company, you are thinking about how can I use AI to drive efficiency across my processes? Can the e-commerce platform itself give me more efficiency and how do at the in the middle of all this change, how do you manage your operational costs? So if we look at the left hand side, all these requests that we are hearing from our customers can be divided into few big buckets. Biggest one is discoverability across LLMs and new channels. Second one, I like to call it agentic merchandising. The next one is B2B acceleration, agentic conversion. And lastly, how can we give our customers agentic development experience and platform efficiencies too? So Shannon in my team actually made it three buckets. She was like, no, four, it is easier to remember. So we are gonna put all these items in three and let us now talk to you first about discoverability. All right, so a ton is happening here.

The crux or in the middle of everything that is happening is first, the catalog itself needs to be exposed. So the cat and of course with your consent with your permission, Adobe is not gonna just expose anything. So the catalog can be exposed to direct feeds to LLMs so that the product shows up. And instead of LLM scrolling the whole internet and being 12 to 13 hours behind, it is much accurate for them to actually query the catalog and provide results from the catalog, which at the same time is far more accurate than crawling that has happened hours, hours, hours before. So how can the catalog be exposed to chat GPT, claw, llama, Gemini, all of that stuff. The second one is there is an entire ecosystem forming that wants to build AI agents and build applications using this data. So an MCP server can actually make all this available. And finally, there is a ton of innovation happening inside Adobe where the catalog actually can help ensure that the catalog along with the content is made visible. And here are certain products that I’ll talk to you shortly about LLM optimizer, brand concierge, site optimizer, experience manager, all of that. Okay, now let us talk about the scenario, which many of you have top of mind for each of you. LLM is emerging as a key channel for you. I have customers who have told me that this is my main channel. So how can we help you there? The first thing is if LLM is your key channel, is your catalog easily and fully and in its best shape and form available to the LLM.

So this is a project in motion. Many of you have this problem. We are taking beta customers. The Q&A window is yours. Tell us how best and my team can see the Q&A. Just put a comment there that you want to be a beta customer for this at no cost to you, but we do need your consent. So if you are willing to be a tryout, like, you know, be a trial for how best your product shows up in LLM and Adobe. Yes, I want to see what’s going on and I want to be a part of what’s going on. Put your contact in the Q&A window and we will be in touch with you. So the first thing is a LLM optimizer project. What does LLM optimizer do? This is where our entire investment in edge comes to our advantage.

Adobe, you remember we have edge based storefront, right? Commerce storefront powered by edge delivery. AEM is on edge. Like many of Adobe’s products are served that way. What advantage it gives us is we now know traffic. We know traffic which others do not know.

So we have that information. With that information, we know what are the common prompts and we can also then figure out how effective are those most common prompts, right? So LLM optimizer helps you identify the problems why your content may not be showing up. First, you make suggestions so that your content does show up and it can go and make those fixes such that it happens. There is no reason why it needs to be limited to content. It can do exact same things for the catalog. We can tell you the reasons why your catalog is not showing up one by one by one. Second, we can make suggestions on what is needed. And third is now I will tell you in rest of the slides how we can automatically go and make some of those fixes for you. So super exciting. This is something like has tremendous momentum. Go ahead, give us your contact information. We are taking better customers here.

The next thing, many of you, yes, LLMs are becoming a primary channel, important channel, but your own storefront. Who wants to lose control of the channel you own? That is your voice. You have full control there. So many of our key customers say my storefront will be my primary channel and I want to ensure it stays that way. So what do I need to do in this world of AI to ensure that I, whoever comes on my storefront, my webpage gets the best experience, gets what they want, gets it quickly. And they keep coming again and again and again. The biggest thing here is need for a conversational interface. We really do not want to go and say I want a blue t-shirt, then I want a size medium, then I want light blue, then I want this kind that like 50 filters. No, that that day has gone. That time has gone. I want to say I’m going to a party. The theme of the party is this, and this is what I look like, give me a blue t-shirt. And I do want to see your best products in the price range I want. And all of you want to give that experience. Once again, brand concierge, we are taking beta customers. If you are interested, go ahead, say in the Q&A window, tell us your contact, we will be in touch with you.

Right hand side, once again, you want to give your customers the best experience on your storefront because that is your channel. You control it. To do that, the storefront cannot have glitches, which means that site optimization agent, that is an agent that identifies a whole bunch of issues. This is slow, this is broken, this is not ideal, and it will identify it, it will make suggestions, it will fix it. And while this works for content and many other things in the context of commerce, it is going to work for a storefront too. So three big projects, LLM as an emerging channel, we can help you ensure your product show up, your storefront as the primary channel, we can help you ensure you have a conversational interface and that your storefront is at its prime. Okay, little bit more on LLM optimizer, right? How do you ensure your catalog and product show up in the queries? It’s not just show up, it is continuous to show up because what do you and I do in LLM interface? We ask 20 questions, right? We just endlessly keep on asking and the answer if it is consistent and if your product is the right one, it needs to come in the answer again and again and again and again and again and again, right? How do we make that happen? So I told you how Adobe knows traffic. We know what prompts are working. We also know what prompts are common. We also know what prompts your product is not really faring well. So now going into the depth of it, I’m not going to do the full architecture. I just tried to make it something very simple. The first party data, even for LLM is your own data, your own storefront, your own catalog, right? Adobe is already talking about this public in the context of content. The same technology now extends to catalog, right? So how can your own catalog be analyzed and ensured? Is it showing up? Is it not showing up? Why? Then comes the third party catalog where you are being talked about in 50 other places. What is important that you or your product or your catalog be talked about consistently in the same manner, right? Everywhere. And that we have found works. So what can we do? An agentic solution which enriches your catalog, right? And I’ll take you describe that in the next slide, what it means to show up again and again. And finally, and also an agentic solution that helps you publish and make use of all different channels. So back to the first agentic solution, which we call it Catalog MeriData Agent, which will ensure that your catalog is rich enough. Now it’s best understood with an example. Let’s say we want to buy a TV. So we’ll say, you know, we’ll say in Chajapiri, for example, top five TVs on sale in United States. Do we stop there? No. Then we say, which one is great for a home theater experience? Then we say, which one will support gaming console? Which one supports a coaxial cable? What is the dimension of this? What are these colors like? Like it doesn’t end. So if you are a Samsung, what do you want? Samsung needs to appear again and again and again and again, even for our 10th answer, which means your catalog cannot just have bare minimal information about your TV. It needs to have information about all these common questions that are being asked, right? So it needs to have all the MeriData, all the titles, all the descriptions, all the tags, all of that, which this AI agent will go and populate in your catalog.

So that is the first part. The second part is such that, you know, all the geos and everything, the layered queries that happens, it continues to show up. Now remember, there is a personalized component to it. Let us say this is United States market where we all have huge homes, for example, and how does it give a theater experience is going to be a common prompt. And let us say we are all in Japan where let’s say the gaming is a common trend and what kind of gaming consoles work is going to be a more common prompt. So depending on the different kind of audiences that are being served, once again, the kind of catalog enrichment that you want to personalize the results is also going to differ and we can help with all of that. Now this is a small thing. Our search and recommendations, as you all know, for many, many years is this AI driven. One of the other things we are doing is that in every channel, it is actually you can do search within that channel and not the holistic whole solution space. This just makes results faster, much more tailored, all of that. Moving on to agentic merchandising. And folks, I am moving fast, but feel free to ask questions in the Q&A window. My team is here, we’ll help answer them. The second one is your channels are growing, right? Earlier, okay, let’s say it was your stores at your storefront, some of your retailers, or you know, like, you know, different, you know what I’m talking about. But what’s happening now is Anthropic, Charge GPT, Perplexory, these are all new channels that are emerging. And how do we actually ensure we can optimise each of these channels, like which product should be emphasised in which channel, and you may choose to send certain products to certain channels and certain products you may choose not to, because there is competition and you like when you use Charge GPT, we see how direct the comparison is. So to avoid undesired pricing pressures and undesired competition and to maximise our own revenues, we know that customers want to do certain products in certain channels and therefore we will have that multi-channel publishing agent as well. And over time, we will do more with it. Now, expansion, all of our customers are here to grow their business, right? Expansion can happen in many ways. Now, 2025, at Summit we released, and prior to that we had many beta customers, we released this new catalogue model, which is once again serving as an advantage. This new catalogue model actually is a four dimensional catalogue where there is only one master catalogue and you can have any number of derived catalogues with the inventory you want, any number of storefronts, which means any number of channels with the inventory you want, with the pricing you want. So inventory and pricing unique to a specific channel or a specific storefront, right? Do it as much as you want. That catalogue now in this world of AI is serving to be a unique advantage that Adobe can bring to our customers. What we want to bring to you now is an expansion solution, agentic expansion solution. Let’s say you are in United States, you want to go to Spain, right? An English language to Spanish language. We can now do that localisation or translation of the storefront of the catalogue for you, right? And you decide, okay, what price you want to sell in that market in Spain.

But this is, this is, people who do it, you know, many of you do it, you know, it’s months of effort, but now first expansion into new geographies, especially new languages. Now the time is shrunk. I like to call it Project Atlas because the vision I have is to shrink the world for you. So this can now happen in days. Yes, there are certain other regulatory things that you will have to deal with, but the things that technology can solve right away will be done for you. In a P2B scenario, so many of you expand into new distributors, new channels, new resellers, right? All of that also takes time and effort and is painful for you, which can be made extremely simple. So this is one example. For example, let’s say we were talking about Samsung, so Samsung itself. Here is an example of a storefront and I actually pulled it off the internet, the storefront on the left hand side in chat GPT, I’m asking about, you know, Samsung, it tells me some of these things. So the storefront, I shouldn’t say PDP in chat GPT and catalog, right? In English and instantly across your entire millions of SKUs of catalog, we can give you the whole Spanish conversion.

Okay. All right. Both storefront and catalog. Agentic developer experiences. I’ll take only a minute because Matt is going to go into this in detail. So we are Adobe and when Adobe commerce speaks to you, we are a small part of large Adobe. At Adobe level, we imagine a world where there will be thousands and thousands and thousands of agents working across complex paradigms and domains. And when that world arrives, there will be a need for a reasoning agent, a brain that can actually orchestrate and ensure these millions of agents work the right way. We are working towards that world from today. A preview of that is the agent orchestrator, which is already, you know, a version of that is out, but commerce will be, and is a part of it. The next thing is the whole, which is top big thing that is top of mind for you. I have so many customizations and extensions. How do I adopt your new SaaS solution? Matt will talk to you about it in detail. I’ll speak just a sec. We are investing in an agentic AI solution that will take your PHP applications and convert it to app builder compatible JavaScript format. You will do 90% of the job, rest, you know, rest does not scare you. The last part, a big investment in MCP servers. You want to build AI agents, the world want to build AI agents and the data now, commerce data needs to be exposed so that that ecosystem can flourish. Yes, we are in that MCP server is coming. It’s coming very quickly where all the commerce objects can be easily understood. The second part is all the documentation, API references, code, and an MCP server, which can make all of that visible as well. Well, that’s it for this session on Business Growth with AI Innovation. Be sure to catch the other two videos in this series for Business Growth with Adobe Commerce here on Experience League.