[Music] [Surya Lamech] Good morning, everyone, and welcome to Adobe Commerce Rockstar Showcase. Thank you for coming in, last day of Summit.

Before we get started, if you see this QR code over here, do scan it and it'll open up the Rockstar playlist. If you have Spotify, add your favorite song to it. It doesn't matter the genre. It doesn't have to be rock. Whatever you like, just add it in over there. We like to keep it as a memory from this session. I'm Surya Lamech, I'm a Product Marketing Manager and one of your host for this evening. [Yasaman Rajaeian] And I'm Yasaman Rajaeian, I'm a Program Manager for Adobe Commerce and your other host for the evening. And before we forget, we tried to get Shaq to DJ this event. He wasn't available. So we do have Rohan who is behind the screen, DJ Rohan. So let's give it up for him. All right, Yas, what's Rockstar about? Yeah. So why we're all here today? What the Rockstar is? So if you've been in the commerce world for a while or not, one big thing, that drives value is for our customers is, like, the community. And whether that's customers themselves or our partners who are doing really creative, innovative things with our platform and the flexibility that it provides. And so the Rockstar is an opportunity for us to really show off some of those innovations from our customers or our partners and show off that technical innovation with Adobe Commerce. And it's an opportunity to really celebrate and show off and share the really cool and interesting things that everyone is doing that is driving business value for their customers. So we're excited to celebrate. So we had 65 submissions this year, pretty much similar to last year. So I really want to thank everyone who submitted. If you-- Everyone who participated in the session. If you are in this room, thank you so much for all your work. We really appreciate everything you're doing for our customers, and all of the great submissions you bring back to us. They covered a wide range of capabilities from back-office integrations to personalization, GenAI, great shopping experiences. So we had a really wide range of submissions this year. And some of those are using technologies that, and submitted for technologies that were still in beta at the time. So they're really early adopters as well, and you'll see some of those innovations on stage. Now to help us here, yes? We have our amazing panel here today. We have Shannon Hane. Shannon, you want to wave? She's our Senior Product Marketing Manager at Adobe Commerce. We have Nishant Kapoor, who is our Director of Product Management, and we have Nicole Cornelson who is our VP of Software Development. And they are here to ask the really hard-hitting questions and help us pick our final rock star. So with that, I'd like to introduce our first participant. So Yaroslav, he's with Atwix. His genre for today is going to be around exceptional experiences. The submission is called Luma Bridge. His favorite artist and I've not heard of them before, but his favorite artist is Rain Geometry. So thanks for introducing us to a new artist and come on up, Yaroslav. [Yaroslav Rogoza] Morning, everyone. How everybody doing? - Good? - [Woman] Great. - After the party? - [Woman] Yeah. Still good? Great. So my name is Yaroslav. I'm serving as a CTO at Adwix. And it happened that at nighttime, I'm musician, sound producer, making soundtracks for movies and Indie Games. And at the day and time, I'm technology geek, business technology geek. And together with Adwix, one of our mission, our goals is to help customer, to help business to optimize the technology and optimize technology in many ways. So first of all, it's performance, then it's cost efficiency and cost of maintenance and many other factors. And talking about the performance, the first thing that usually come to mind for B2C, not only B2C, but mostly B2C business, is, the storefront, basically. To have a good customer experience, you need fast and reliable storefront. And fortunately or unfortunately, I don't know because in Magento, sorry, Adobe Commerce, we have the storefront, we have Luma-based storefront, blank-based storefront, but they aren't far from perfect, unfortunately. So usually when we are talking about improvements, it means go into some alternative solution and for many times it's headless or edge-delivery services or similar storefronts. But the main problem here is going from the standard Adobe Commerce Luma storefront to headless. It's not just creating a design. That's replatforming because if you build a lot of customizations around your Adobe Commerce store, you need to do the same for your new storefront. And it's not like a migration. It's, for most cases it's creating solutions from scratch. So what we did is we created a bridge between, like, your current storefront and headless or whatever you like. So you don't need, like, your business don't need to spend like huge budget for this replatforming and plan it for two years, three years, making it a huge project. So with Luma Bridge, this solution that I'm going to present today, you can do it in portions, basically. If you want to migrate your store to headless, you can do it page by page. Start, for example, from homepage, having homepage on headless and everything else on your Adobe Commerce standard storefront. Then when you are ready, you can, like, develop the product detail page, product listing page on the headless part and leave the checkout on the Adobe Commerce part, for example. So, basically, it's one of the business cases of the Luma Bridge. And here is description, short description diagram of how it works. So basically, on the upper part, you see the headless storefront or PWA or, like, storefront you like. It might be custom-built storefront, and the lower part is standard Adobe Commerce part. So you can perform different actions on your storefront and then redirect customer to checkout or to any other page of your Adobe Commerce website. And moreover, it's not only the redirect, it's the transparent transition with the same URL, same design, so customer even doesn't know that he uses two websites. So the Luma Bridge allows you to use both, of, like, best of two worlds using headless and using your standard storefront with your own customizations until you're ready to migrate everything to headless. So I'm going to show a short demo how it works. And it will be a little bit tough because-- Yeah, I see this. Yeah. Okay. So now on the screen, you see the alternative storefront. Let's call it like that. It's built using Edge Delivery Services with drop-ins. If you're not familiar with Edge Delivery Services, feel free to find me or a GogoDancer that are dressed in red kimono. So pretty easy, and we will tell you about the solution more. But, on this demo, you can see that you have the homepage. It's like built on the Edge Delivery Service. You have-- Let's go to some product listing page. So you see the product listing page with, like, layered navigation, everything else. And if you go, for example, to product detail page, you can add your product to the shopping cart. And just want to emphasize that everything happens currently on the headless part or on the EDS part in this case. So we add that product to the shopping cart. We can see the shopping cart contents here. And like we-- Let's imagine the case. We didn't have much time to migrate our checkout to EDS because as a rule, it takes a lot of time. If you install some extensions from marketplace to your Adobe Commerce store for payment providers, for shipping providers, you, most likely you want to keep it till the end because I would say that's the most challenging part to migrate the checkout. And we currently can click on the cart page.

And as we see, we are redirected to the standard Adobe Commerce theme, but the URL stays the same and design everything else the same. So it's not, like, visible for customer that it's a different website. But you can see that on Adobe Commerce site, you see that this product you have added previously to the shopping cart. It's here. You can go to checkout. You can perform the checkout with this product here. And once you are done, you will be redirected back to your storefront, like this EDS storefront without changing URL, changing design, etcetera. So let me, maybe check a couple of examples. So let's go back to the cart page. This cart page is on Adobe Commerce site. And, if we click, for example, any link in the footer, we will go to our EDS part with the same URL, with the same design, and everything else. So just to wrap up, I would like to show more examples, but I don't have much time, unfortunately. So if you are interested in more examples, please find us. And also just wanted to demonstrate maybe how it works on the technical side of things.

Just a second.

So it's just a brief diagram how it works. So you don't need some enterprise level system to use the Luma Bridge. Basically, all the routing, all the, like, decision where customer should land on the EDS side or Luma Bridge is performed on the CDN. And you don't need to purchase some, like, some enterprise level CDN solution. You can use free tier Cloudflare Fastly. So it's possible. We use workers for directing the customer, directing the traffic to the correct point.

So that's basically it. As I said, I would like to show more examples to you, but let's leave some time for questions. So thank you.

All right. Thank you, Yaroslav. We have one minute for questions from our panel. We have a question for Yaroslav? [Man] I have one question. Can you show page speed on the homepage? Can you repeat this? Show Lighthouse score on the homepage? - Page speed score? - Yes. Okay. Let's try to do that. It might take some time.

But, sorry. It's not my laptop. This link is in my bookmarks as a rule because I use it every day.

Wow.

That's fantastic. Cool. Thank you. All right. We have our next Rockstar that I'm really excited to introduce. We have Shikha Raina. Shikha joins us from Bounteous. Her genre of her presentation today is around back-office integrations and the latest release of Lean Order Management. And her favorite artist, who I'm also a big fan of, is the Imagine Dragons. So round of applause. Shikha, come on up. [Shikha Raina] Hi, everyone. I'm Shikha. Today, representing Bounteous times Accolite. I'm delighted to introduce our latest innovation, leanest OMS. This solution has been built for efficient order fulfillment through instant messaging services.

So one of our existing clients came to with some challenges they were facing in a different geographic location. And the business was running on a proprietary platform before. So the business operated, like the model was like the, they had warehouses, and each warehouse had a certain set of drivers assigned to it, and they were selling the consumer goods from the warehouse.

So let's look into the core issues they were having with that platform. So since that platform, that platform was having consumer website and native mobile app for communication between the warehouse and its drivers. The main challenge for them was extensibility. They were not able to extend that platform with the latest market trends. And it was becoming extremely challenging for them. Next was the issue with the manual processes. So the communication between the warehouse and its drivers was manual. Assigning the orders to the nearest warehouse was manual. All these created bottlenecks for them. And then they wanted to achieve same day deliveries to enhance their customer base. So that was not getting efficiently done. So, and the last challenge was cost. Extending the proprietary platform, making customizations on that, making new integrations with the new services in the market, it was all challenging for them.

So when they approached bounty as we gathered our commerce team, engineering team, delivery team, and we came up with the best solution for them. Let's look into the solution in detail. The first step we did, we knew that they had tremendous growth aspirations. So the first thing we did was migrating them to the Adobe Commerce Cloud.

Because this is the platform which is designed for scalability, feature enhancement, and optimal performance. The next thing we did was, the next problem we solved for them was the manual processes. We used the distance priority algorithm, which is the out of box feature of Adobe Commerce, to find the nearest warehouse between the shipping address and the warehouses. And the third one was efficiency. To meet that, we integrated with WhatsApp. WhatsApp was readily available on the end-user mobile devices. And we integrated WhatsApp with App Builder and Adobe I/O Events. And it was a leading platform, instant messaging platform in that, region. The last was cost. Since WhatsApp was there already on the users' mobile devices, we didn't have to build any native apps to support this integration. And this resulted in significant reduction in the costs.

This is our architecture diagram. On the left side, the blue box is our Adobe Commerce website. With each order in the system which comes into the Adobe Commerce system, there's a notification triggered on, to the, through Adobe I/O events.

It's all in-built like, Adobe Commerce and I/O events are already there. It's available. And once the order comes, that event is triggered, and the relevant data goes to the App Builder application, from App Builder application it is programmed to trigger WhatsApp Business API, and the message gets triggered to the end user mobile devices. So, once the user takes action on the WhatsApp, the response of that user action comes to the API Webhook, which is within Adobe Commerce and the order gets updated. And the Webhook is the place where it decides what action is next in the workflow.

Okay. This is-- Let's see all these components in action. This is our demo. So our integration starts once the order is placed into the Adobe Commerce. Once the order gets placed and the order moves through processing status, that is the invoice is created and the order is paid, that's when our system gets triggered. So the next step is to determine which is the closest warehouse. So all the calculations are happening in the backend once the order is moved to processing status. So you can see there's awaiting warehouse assignment.

This data goes to the app builder. From app builder, we will trigger a WhatsApp message. For the purpose of the demo, we have used WhatsApp web, since it was not possible to show the mobile application here. So this is the warehouse. It has accepted the order. So once the status has updated in Adobe Commerce, warehouse is assigned, shipment is created. Basically, it's an MSI. The source is selected. Now, the next step is to choose driver for the order. So what we will do, if the warehouse has one or more drivers signed, you will get a list populated in WhatsApp to choose from the driver, to choose the driver.

This response action goes back to commerce.

It will move to ready for pickup state. Once it moves to the pickup state, the next part is to notify the driver. The driver will get a WhatsApp notification, that please proceed to the warehouse. Pick up this order.

This is the notification for the driver.

The next step in the workflow is to mark the order as picked. Once this is done, the warehouse gets another action item to do. That is to mark the order as collected or picked. Once this is done...

The order gets updated. At this point, we can do more things if you want, like notify the customer that your order is on the way and stuff.

Then this is, the next message is the instructions.

This template is entirely customizable. It depends on business discretion, what data you want to pass to the driver, what you don't want to pass. So it's all customizable in the WhatsApp APIs. And the last step is now to mark the order as completed.

So for the purpose of this demo, we again assign that action item to the warehouse and warehouse can mark the order as completed. Once they do this, the order gets updated in the Adobe Commerce and it's completed.

For the ease of the demo, I used one phone number for all the warehouse and drivers, one WhatsApp number.

So these are screenshots of the messages just in case we want to see them in detail. The first one was the action item to send where the warehouse accepts or rejects the order. The second one was to choose the driver from the list of assigned drivers. The third one is notification to the driver to proceed to warehouse. The fourth one is when the driver collects the order, warehouse can mark the order as collected. Fifth one is the delivery instructions for the driver. Sixth one is to mark the order as completed. This is our App Builder Event Logs screenshots, configuration screenshots, how we configured and just to give a sneak peek.

Now, Benefits and Impact. So the first one is Accelerated Development.

This solution, we had to write very minimal code. And the app builder, since it is a low, it's a low code and no code platform, everything was readily available, the events we wanted to hook onto. Even we didn't have to acquire any new skill set. Everything was there within our Adobe Commerce practice. This accelerated the whole project timeline to another level. Next one, the ecosystem. Since the client came to the Adobe Commerce Ecosystem, they were able to leverage a lot of commerce features which are not there for them in the previous platform. And they were able to tap into the customer behaviors through AEP and give personalized offer to the customers as well.

Failover Mechanism.

We have failover mechanisms auto configured in App Builder and Adobe I/O Events. And it can be configured by the user as and when needed. So this enhanced the reliability of the whole application.

Cost, of course. So there was no need to integrate into a cumbersome order management system, no need to create a separate mobile app for the warehouse and the drivers. So it was really cost effective. Scalability, since day one they were on Adobe Commerce Cloud, within a year, they experienced tremendous growth. They did not have to change platforms or think about how we will go from here. So it was all good for them from the starting.

The extensibility. So all these components you saw in architectural diagram are loosely coupled. They don't have any much interdependencies on each other. So whenever the new market technologies come, you can replace any component with ease and without much impact to the other component.

Last but not the least, a huge shout out to the Bounteous team who made this rockstar submission possible and this project a success. Thank you everyone for your hard work and dedication.

[Woman] Hopefully, okay. Hi. That was an excellent demonstration. Thank you. I am curious if the client has come back and asked for additional sort of out of process extensibility features using app builder. - Do you have any plans to do more? - Yeah. They keep coming to integrate various other Adobe ecosystem platforms. Like, they came after one year to tap into the customer behavior, AEP. Now they have plans of doing CDP and everything. - So-- - Okay. They keep on coming to us. All right. Well, good job. Yeah. One quick question. [Woman] One quick question. Can you compare how long it would take you to develop using App Builder and this new extensibility framework versus what you may be used to do in the past for something like this? So for the past, doing something similar in RapidMQ when we did it, we took around two and half to three months. And this one, we did it in few weeks.

All right. Thank you, Shikha. That was amazing. I love that demo.

So thank you for doing that. Next up-- Next up, we have Martin. Martin is from Comwrap Reply. His genre is GenerativeAI, and his submission this time is the Master GPT Sales Component. His favorite band is Snarky Puppy, which, again I learned so much from these sessions, not just technology, but music as well. So thank you, Martin. Please come on up. [Martin Altman] All right. Thank you. By the way, this was not a Snarky Puppy song. Snarky Puppy is way more chilled. So my name is Martin. I'm from Comwrap Reply. We are Adobe Platinum Partner from Frankfurt, Germany. And I'm going to show you the Master GPT Sales Component, which is not the perfect name, though, but didn't have a better one. So it's an app builder application at its core that utilizes content that's generated and validated by OpenAI. This content is based on suggestions from Adobe Commerce product recommendations to ensure relevance to boost sales. So what does it actually do? In the end, it's just very, very simple. The tool-- Wait a second. Yes. Just takes boring product data in and exciting product data out, right? So we created this together with our client Sunstar here for the Canadian market. I will come to that in a second. But in the end, this is what you will get, right? It's not super exciting, but, actually it is super exciting in the back, what's happening here. So you get a product teaser or even a later release, a complete product landing page based on OpenAI data. So, sorry. So, again, exactly for whom did we build it? So first of all for teams that are heavily under the gun, meaning small marketing teams that are under heavy pressure or even big marketing teams that are under heavy pressure. Also, we built this for product that needs a lot of attention and a lot of product that need a lot of attention and sorry. Quickly get back. And for a faster selling off, so for situations where you really quickly have to sell off stuff, right, for reasons. So how does it actually work? It's the Master GPT sales component, first, what it does, it does a product so an API call via the app builder to the recommendations API. So I wrote down here what the recommendation API can actually do. It's most viewed, most purchased, most added to cart trending, yadda-yadda-yadda. And actually, one request we had is we need the least trending product because this is also very interesting for us, right? So next thing is once we found the SKU, let's say the most trending product, we query via the-- We are the GraphQL API. We query the product Metadata, and we take the boring Metadata, right, and send it over to OpenAI with a prompt that's just create a great teaser, right? We use the JSON mode for it. We'll come to this maybe later, which is quite interesting. And then actually the most amazing part for me is happening here is that we don't publish the content right away, what we're getting back from OpenAI. We validate it with an n number of prompts again. And I will come to this in a second. This is actually what we found here it's quite interesting. So once validated, right, we publish the content. And if the quorum let's say, the validation says it's crap, we just do it again, right? I think we do it again for three times. I'm not really sure. Yes. So, these are all the toolset. We use one thing. I guess you know most of it. The most important part is the Adobe Experience Manager here is not relevant, right? This is important, right? The most relevant part is that we have the App Builder and Adobe Commerce here. So for us, in the demo, we show in a second, we use AEM as a frontend. This is the architecture. So in this example, you go into the Adobe Experience Manager. So you select your homepage. On the homepage, you put this master GBT sales component as a component on. You configure it. Let's say, take the most trending product. Then it goes down to the app builder. From the app builder, it goes up to the recommendations engine. It gets the SKU. It takes Adobe Commerce to get the product data out. It sends it then to OpenAI. It's the fifth step there. Gets amazing product data, like amazing teaser information out, validates this via three agents currently, and then sends us back to the website, right, to publish it. This is another example, which is important. The core is actually the App Builder application. That's it. You can replace recommendation service easily. You can replace your Adobe Commerce if you want with any PIM. The only thing that sticks there is, for now, OpenAI. We use OpenAI via Azure for compliance reasons, not directly OpenAI. And even the storefront upfront is like, you can take anything, your techy react application, even take WordPress if you like, if you wanted to, right? And so-- And one thing I would like to deep, a little deeper in-- Sorry, dive a little deeper in is exactly this. So this is the API call we are doing when we do the proofreading.

Two things I would like you to focus on. The first part is very funny. If you tip your AI, it will give you better responses, right? And $200 is obviously the best thing. There was a research going on, and we added, and it was actually better, right? And then the two other things that's important, always take response form at JSON Object. You will always get a valid JSON back way easier, right, to ingest it. The second thing is set your temperature not to one, but zero. Very close to zero because then OpenAI will not hallucinate in your answers because this is obviously very important. So, this is the call we did, and what I actually expected at firsthand was, okay, send content, the same content over. The first prompt would say, so you see, I'm asking for a score xxx. So I'm asking for a score between one, uh, zero and one. And I was expecting that. The first prompt will return zero to five, and the second prompt of the same content will return 0.7. It will be just hallucinating all the way, and it did not. This is the craziest part for us. So we did we did create some, we did create a diverse teaser content, product teaser content, so titles, subtitles, descriptions, and stuff, and sent it over. We created a really bad one, right? Somewhat medium description, a medium one, and a really good one. And we did 150 runs of the same content against OpenAI. And what we always got back was the bad one was always 0.25 and medium one always 0.45. Sorry. Somewhat medium 0.45, medium 0.65. And we never ever, by the way, got over 0.85 for a reason, but we didn't really care. So by this, we normalized the whole thing. So we know 0.85 is a really good teaser. So that means we run it 150 times, same content. It was always 0.85. And this is amazing because when you think about it, so if it's not about life and death or, like a cancer diagnosis, you can actually validate your content with OpenAI, which is like in content workflows where we listen to all the talks from Shantu. Sorry, Shantanu, who says, like, assets will explode, content will explode. Like, who should review all this, right? At least the one that's not super important, let's say not regulated. So if there's little mistake, then nobody will care, right? So you can actually use OpenAI to validate your content. So nearly last is demo. Is it running? Now it is, right? So we're in AEM. We place the master GPT sales component in. We configure it. We currently, we don't use most view field. We actually use a just custom selection of a product. We pick the product, and we salute, we choose what we want to have back from OpenAI. We want to have the title description. The image by the way is not coming from OpenAI, but the rest is. We pick a sales quantity and how many days it should run. Sales quantity means how much it should sell before we replace it with a new one. So we can automatically actually sell it off, next one. Sell it off, next one. And this is it, right? Is it there? It is there. Yes.

So the conclusions is, I don't need to do, like, guided reading here. I think the most important bit is that we do optimize product recommendations, right? Really fast paced product recommendation. We can really boost sales. We have perfectly integrate this in Adobe ecosystem, so everyone with the app builder can start using it.

And I think the most important part is this consistent quality that we can create during with this quorum validation. And what we actually learned, I think I said this already, is that for me the most important, so the most, I'd say astonishing thing was this validation bit. Then we actually use the validation part for our internal merge requests right now. So when we merge code, we not fully, but we automate like a quite a bit of it. So what I want you to just let you know, like, be creative, right? Think outside the box. You have with the app builder, actually, some kind of, like, crazy middleware with a database. You have everything in there. You can connect. Easily you can connect services and build something like we did. And the last thing is, like, tip your GPT, like, tip your OpenAI, right? And the biggest thank here definitely goes out to Alex here. So, actually, we created this together.

We created this together. We sat down a lot and draw a lot, and then Alex coded, actually all of it, right? And so when we have technical questions, right? And actually, we rolled the dice and I lost, so I'm on stage, right? So, and there is-- By the way if you can scan this code, it's a technical walk through from Alex, and we open sourced all this so we can download it there as well. - Correct? - Yeah. - That's the link there. - Yes. So have fun, like go in there and download it. And thanks a lot. That's it.

All right. Thank you, Martin. Do you have questions from the panel? Yeah. Hey, I'm curious to know, have you actually put this into market and how, or is this still kind of a proof of concept? No, it's not a proof of concept anymore. That's on UAT currently. Right? So we're still testing it. We are still talking to the Canadian fellas and, but we are on the cusp, right, doing it. Okay. Yeah. No. They're American, but it's for the Canadian market. But anyhow, right? Well, then quick follow-up question. What else do you have in your bag of tricks for using, ChatGPT or Open AI kind of-- Come again. What else are you looking to develop next? - What else features? - Yeah. So obviously-- Yes, everyone would say, yeah, use Open Source LLMs connectivity. Here I need to say we tried. It's not the same. Or like, GPT 4 is just so much better, even than LAMA 2, Baklava, all these multimodal languages, sorry, models. Even better than Gemini still I feel. So, currently, for now we will stay with OpenAI via Azure because for us at least the most best and most regulated, so in the in the midst of things. And we start actually with creating a full landing page. So this would be the next feature. We started actually with it and then I thought it's a lot to ask for GPT. It took a while to create the full text for a language page, even though it worked. But this could be a next step to maybe separate the actions into three, four prompts, right? One prompt for the hero, one prompt for this, this, that. So you combine in the end, what you get back from OpenAI. Maybe you validate that and then put it online. So my goal is actually create a full-fledged landing page, not just a teaser. But we thought this is the most tactical use case in a way, right? It's most efficient and less work. Yes. - Thank you so much, Martin. - Sure. All right. Another round of applause.

All right, panel. Now you have a tough decision, so we're going to send you either in your corner, if you want to leave the room to deliberate. All right. Well, first off, I just want to say thank you for all of our finalists here. Everything was amazing and I think, just really incredible work happening. And also the thing we noted is that, so much of the newer technology that we've brought to market was actually put to work here today and it's really gratifying to see that and see what you can do with it. So I'd like to announce one of our runner ups and that's Yaroslav. So let's give a big-- And just a couple of things that we noted on the judging panel. First off, we really rolled the dice, and we showed the Lighthouse score. And you could see the incredible performance of Edge Delivery Services at play there. But we also know that our customers don't have the luxury of just flipping their storefront overnight and moving to an entirely different solution. And so the fact that they can have this bridge between what they have today and where they want to go, which is to a faster performing storefront is incredible. And I think we'll see lots of customers being able to take advantage of this capability, progressively moving to a better storefront. So in terms of, kind of impact to customer, helping with affordability and maintenance, things like that, it's right on the money. So thank you so much.

Thank you, Yaroslav. All right.

All right. So I'm going to announce the second runner-up, which is Comwrap Reply. Round of applause.

What an incredible app. It's solving the real-world problem. And there are merchandisers always under the gun. My friend, Shannon, she could use that for sure. Absolutely. So fantastic, the solution showed the flexibility of your design to plug-in different systems. And again, thank you for believing in App Builder and leveraging that for your solution. Great work.

All right. Thank you, Martin.

Okay. So, wow. That comes to our winner. And our winner is Shikha from Bounteous. We are so excited about that app.

Very, very nicely done. Happy to see that it's working for the customer and in production. That's always a great thing. We really loved how it integrated with some of our core algorithms from commerce. We love that you used App Builder. It's a very great thing for all of us when we have partners getting that value for customers as fast as possible, maintainable, easy to switch out. We want to go towards this model with Adobe Commerce, loosely coupled, easy to change things out, and App Builder obviously is a great way to accomplish that. So congratulations, Shikha. Good job. Great presentation. Thank you, Shikha. Congratulations. All right. Thank you everyone for the session and thank you to our participants, most importantly. So thank you so much. We'll enjoy those demos. Love them. One more round of applause for everyone. Thank you. [Music]