AEM GEMs - Maximize the impact of your web experience with AEM Sites Optimizer

This session will take you under the hood of Sites Optimizer, an intelligent, AI-first application designed to enhance the performance and impact of enterprise websites. Learn how it leverages powerful agent-based automation to continuously monitor site performance, identify friction points in traffic acquisition and user engagement, and deliver actionable recommendations. We’ll demonstrate how it enables marketing and development teams to proactively fix issues, improve SEO, accelerate page speed, and enhance user journeys — all while aligning with business objectives. Whether you’re a developer, marketer, or digital strategist, this session will show you how Sites Optimizer can become a critical tool in driving measurable outcomes for your web experiences.

Transcript

All right. Kartik, feel free to start your presentation.

Awesome. Thank you, everyone, for joining. I’m just going to pull up my slides. So we have a great session for you today on AEM Sites Optimizer. So hopefully many of you have already heard about Sites Optimizer. For those of you who don’t, no worries. It’s about 3 months ago. We’re already seeing tremendous success with customers and I’m really excited to go deeper into it today and talk about the value prop, show a demo, and then answer all your questions and help you be successful with this product eventually. So I want to first talk a little bit about what this even is. So we, working with many different brands, our AEM team, we know that every brand has this problem. And this problem sits with someone, typically, let’s call her Emily, who’s a marketing manager. But at your orgs, it could be a product owner, product manager, digital experience owner. It’s a business persona, ultimately. And this persona is really concerned about how do I actually bring traffic to my web channel, get that traffic to engage and ultimately convert it. So we know AEM is just a world-class solution in terms of building and delivering these great digital experiences. But it’s another problem in terms of how do you continue to optimize the traffic that’s coming in, make sure they’re actually engaging with the content and ultimately converting. And that’s what Emily here is concerned about. But we know that people like Emily in your organization have many different questions and issues that they’re thinking about on a day-to-day basis. One is, hey, they are not getting enough organic traffic to their pages. Or maybe they have paid campaigns that they’re running, but that paid traffic is not converting at the rate they’re expecting it to. Maybe that traffic is not the problem. Traffic is coming in, but ultimately they’re seeing drop-off or abandonment rates in the middle of the funnel. And that might be due to the content not being really relevant or resonating with visitors. And then ultimately, there may be something that happens at the end of the funnel where they’re just not converting. So they get all the weights, that form fill, or that final transaction, and they bounce.

And so how can we help Emily address some of these issues? And one of the challenges we know and all the customers we talk to is that web is not simple. We’ve been working with the web for years now and many of you are very familiar and can probably point out these different issues that you see on your own sites. This is just a sample list of the types of issues that Emily sees impacting their site on a day-to-day basis. And we see these types of issues for all our customers, whether it’s accessibility issues, whether it’s JavaScript errors, a complete site map. Some of them are very technical, some of them are also less technical and they’re more tied to content is not really being engaged. They’re seeing traffic bounce. So these issues are coming on a regular basis and Emily and her team is constantly firefighting and trying to stay on top of all these issues. And the challenge really is the way they stay on top of these issues is through a variety of tools. So today, as many of you might be also working with, we have site audit tools such as Site Improve, SEO performance tools like SEMrush and Bright Edge, and web performance tools. There’s developer tools they’re using to help optimize the website and web optimization tools. So there are multiple tools that most organizations are using to solve this problem. And each of these tools does something a little bit different. But for the most part, these tools are very good at identifying the problem and figuring out where you’re actually seeing these issues happen. But after the diagnosis, one of the challenges we know that Emily faces and others face is how do we actually implement the fix for this? How do we come up with the right solve? And then go ahead and implement that. Now, these solutions require multiple teams and collaboration across multiple personas. There’s SEO teams focused on the SEO issues, content teams working in creative teams who you might need to ask them to create some new assets that you want to put on your page, and then developer or IT teams that are maybe working on the more technical issues of the site. So ultimately, these are all these different teams and personas that need to get involved. And getting them involved takes time. It’s Jira tickets, it’s lengthy dev cycles, just to fix some of the most basic issues. And so for Emily to solve all those issues we showed a couple slides earlier, it actually is quite time consuming and quite cumbersome to actually get to a resolved path. And this is the problem we’re seeing. And we saw an opportunity to help solve for our customers. And we are struggling to make a meaningful impact because of this orchestration across multiple teams and tools. And ultimately, the big reason this is a problem is it limits revenue growth. And that’s a top line metric that the business cares about and ultimately needs to solve for. So that is where we introduced AMSights Optimizer. And AMSights Optimizer really was all about how do we empower that market or that business user to deliver better business impact by optimizing for that traffic acquisition, the engagement, and then ultimately the conversion on those experiences. And how do we simplify that whole manual workflow, multiple tools, bring it into one application, and use AI use an agentic approach to actually help address this in a simpler and more cohesive way.

So, Sites Optimizer, what are we actually focused on? Really, when we think about that customer funnel and where the drop off is happening, Sites Optimizer is all about how do we ensure we’re expanding every part of the funnel and optimizing it every journey along the funnel. And that starts at the beginning with technical SEO. So we know technical SEO impacts organic traffic, things like site speed, things like your site architecture, core vitals. How do we fix those and help customers optimize for those? But SEO is also tied to content related SEO, where it’s optimizing for non-branded keywords, relevancy. How do we help our customers be successful there as well? There’s also channels like paid media or social or referral channels. How do we actually, when that traffic comes in, how do we make sure the experience they see on the web is consistent with what they’ve seen through those paid campaigns and making sure that they can engage with the content and continue to find it valuable? As we go down to the engagement, this is all about, okay, once you’ve gotten the traffic, how do we make sure the content itself is driving clicks? It’s driving engagement. And if there isn’t content that’s working, how do we replace that with content that is? We know accessibility is top of mind for every single customer. How do we make sure that we are addressing all the accessibility needs, ensuring optimal engagement for all visitors and making sure that you’re meeting those compliance standards? And then finally, conversion. For us, conversion, really, we think about the forms experience, whether that’s a lead gen form, a transaction form, how do we optimize those call to actions? How do we optimize that experience to reduce abandonment? But also thinking about security fixes and things that might be ultimately preventing conversion is something that we care about as well. So Sites Optimizer is not just an SEO tool, it’s not just a content engagement tool, we really are solving for all aspects of the customer funnel to make sure that these visitors are engaging and converting on your site. And what’s exciting about Sites Optimizer is the framework we’re using. And the way the agent actually is doing this behind the scenes is through an auto-identify, auto-suggest, auto-optimize framework. And very shortly, Hyman is going to show this in the demo and what it looks like in practice. But auto-identify is all about using the operational telemetry data we collect for AM customers today. We have a data source around traffic, clicks, and we also are using Google Search Console data. We’re also using other SEO related data, multiple data sources to help identify what are the largest issues on your site, where are the biggest opportunities for you to drive business impact. So the agent is able to actually identify these opportunities, sort them based on priority, and then surface that to the marketer. This is all available out of the box for AM customers, so there’s no implementation needed for this product. We can start doing this on day zero. Now auto-suggest, this is where we can now take it one step further and start providing very prescriptive recommendations on what is that code or content recommendation that’ll help address that opportunity. Now this is using our LLMs that are using your context, your implementation details to give you very prescriptive recommendations that make sense for your brand, incorporating your brand guidelines, so very highly tailored recommendations that you can start to adopt. But then we take it one step further with auto-optimize, and this is where we want a one button execution, one click execution of these recommendations available in the application. So we know that where a lot of the bottleneck happens is the fact that a lot of this is manual today. You have to go to your developer technical resource to go fix these things. Now with Sites Optimizer, all you need to do is click a single button as Hyman will show you, and you can auto-optimize these things immediately with no technical expertise required. Now I talked about agents, but the reality is throughout this process, the marketer, the business owner has oversight and can review the opportunities, tweak the recommendations, and before you click optimize can review what that recommendation and what they’re going to publish is. And there’s also review and approval workflows built in to make sure that other teams, corresponding teams can provide their input as needed.

And as we built this product, we had some distinct guiding principles that we wanted to follow. I mentioned that this is pre-instrumented, so all our AEM customers on the cloud, managed service, edge delivery, they have this operational telemetry data set already, and so we can turn this application on with no heavy implementation project. So you can start seeing value immediately with this product. Customer-led innovation is how we do everything at AEM. We’ve been building this with customers and prioritizing the opportunities that impact the most customers and where we see most customers struggle with. So our roadmap is constantly evolving, we’re constantly adding new capabilities based on where we see our customers struggling the most and what can drive the biggest business impact. User control is important for this product at every step of the way. The user is in control of the automation or can review all the recommendations, review before everything gets published, and ultimately what we want to do with this product is transform marketers into superheroes in their roles and in their organizations.

So this is all about how do we empower that business persona to be more effective at their job, to address more of these issues and tie it to business impact and show that the impact that they’re making with Sites Optimizer is leading to revenue benefit for the organization. So it’s important to make that market or that business persona, or whoever in your team is really responsible for this, making sure that they can become more strategic and they’re not focused on the mundane details of executing these different activities.

And the last thing I want to talk to before maybe I pass it to Hymen is just the the core use cases we talked about. So Sites Optimizer really focused on four different areas. You’ll see a set of capabilities around page performance, so specifically how we optimize for core web vitals, what are the prescriptive code recommendations to get you to a better page performance and to help you maintain that. Traffic acquisition is all about our SEO optimizations, so looking at things like broken backlinks and how do you optimize fix those and optimize for organic traffic. And content engagement is a really exciting one because this is all tied towards content and how we can give you different variations of content that will work better using AI and allow you to test this in a seamless way or publish this in a seamless way. And ultimately conversion is tied to those forms related experiences at the end of the funnel. Actually we’ll stop there so we can get to the demo. So Hymen, I’ll turn it over to you to show what this looks like in practice. Very good. Thanks Karthik. Hi everyone, my name is Hymen Chung. I’m a Product Manager working on AM Sites Optimizer. Very excited to be with you here today. Karthik, I think you need to stop sharing. All right, yep, there you go. All right.

Very good. Okay, so this is AM Sites Optimizer for a fictitious retail brand Frisco Pecafi, but it’s representative of how the technology will work for any AM cloud service, AMS or edge delivery services site. So first thing you’ll see is the different metrics that AM Sites Optimizer is measuring and optimizing for Scopa, that for Scopa site for. You’ll see that we’re measuring and optimizing for traffic acquisition, measured by page views, engagement, measured by clicks on different elements of the page, and then user experience measured by Core Web Vitals. It’s very important to note, as Karthik mentioned earlier, that these metrics and Sites Optimizer doesn’t require a separate implementation or instrumentation project. So as long as your site is on an AM cloud deployment, we’ll be able to start collecting this data and helping identify where there are opportunities for optimization. And then below that is very importantly the different opportunities that Optimizer has detected for the site to drive better business impact. And they can be things like high authority domains linking to invalid site URLs. So if you’re familiar with SEO, this is essentially broken backlinks. The original kind of breakthrough from Google is understand and measure the importance of one site versus another. You count how many other sites are linking to it. So when we get into a situation at Frisco where there are other domains trying to link to the Frisco site but sending to 404 page, they are losing what’s called link equity from the search engines from Google and the other search engines. So when I recover the link equity from the different backlinks that are sending traffic and search engines their way, so that’s the auto identify capability that is in the framework that Karthik mentioned earlier. So identify across the web all the different domains that are trying to link to Frisco but sending to these 404 pages. Then what the agent, the Site Optimizer agent is doing is taking the anchor text of this link as well as the different tokens in this URL to identify the intent of the backlink and then finding the best match to a live page based on its understanding of the different content across the site. So it’s matching content to intent here, providing this suggestion, providing rationale for how it made that match, allowing the user to go in and change the target path if the agent doesn’t get it quite right. Also allowing them to ignore the broken backlink altogether if they choose not to resolve it. So that’s the auto suggestion. And then once the marketer is ready they can click this deploy optimization button to connect these redirects into the right place in their site implementation. If it’s edge delivery services it can be the redirect file, if it’s cloud service or ams it could be in their ECS commons redirect mapper. And that’s now how the marketer can recover from these broken backlinks and drive better organic performance. So they don’t need a SEO tool to try to identify these broken backlinks or analytics tool to find the 404s from referral traffic. They don’t need to spend hours or agency costs to try to figure out how to redirect the traffic. They don’t need to create separate IT Jira tickets into the IT team’s backlog to get these redirects into the right place in their site implementation. They have the solution that does all of this from within Sites Optimizer. So that’s an example of an SEO optimization that’s available within Sites Optimizer. We also have optimization for user experience measured by Core Web Vitals. So we are measuring and identifying when there are regressions in Core Web Vitals across the pages on the site, across device types as well for the different Core Web Vitals metrics. And they’ll be surfaced here when we detect that there is a page that is not in green Core Web Vitals. And then we’ve trained a model to evaluate the site implementation when there is this regret these regressions in Core Web Vitals and provide very prescriptive down to code level instructions for how to recover and address the issue from a code perspective. So whether it be cumulative layout shift or largest contentful paint issues, you can see the very prescriptive, very specific instructions for the technical teams to fix the issue that is causing the regression Core Web Vitals. We know there are other tools in the market like Lighthouse, like Google Search Console, that will give you generic recommendations on generally what to do to get to green Core Web Vitals. The difference here is the level of specificity that’s included, very specific to AM implementations, very specific to the specific site project and how it’s been implemented and from these issues. And ultimately the engineering team will get a pull request with these enhancements and these optimizations so that they can review and merge this optimization into their code repos. So that’s Core Web Vitals optimization. We also have optimizations for accessibility. We know it’s very important for all types of visitors to be able to consume and engage content on the site. We also know in the EU there’s new regulation that requires accessibility to WCAG standards AA, otherwise they potentially could be fine. So we want to of course help our customers get ahead of these types of issues. And one way that we’re doing through Sites Optimizer, identifying when there are things like missing alt text for images across the site. So here what we see is the auto identify capability surfacing the images that are on the site that are missing alt text. Then the auto suggestion is using the latest vision model, evaluating the image and providing a suggestion for what the alt text should be. Again, the user can come in here and edit the caption if the agent doesn’t get it quite right.

And again, once the marketer is ready, they can click the deploy optimization button and this will write this content into the right place in their content repository. So it’s important to note that Sites Optimizer is very tightly integrated into AM and has APIs into the right place in the content repository. So everything, whether it be code or content goes keeps your site implementations clean and supportable and maintainable going forward. So we’re not doing anything where content or code gets rewritten post render time. This is all supportable and maintainable code and content in the right places following best practices.

So that’s alt text optimization. One additional, I’m not going to go through all of these, but another really interesting optimization of course is for engagement. So we have an agent that detects when there are opportunities to improve engagement and bounce rates for different experiences on the site. So the auto identify capabilities here is detecting when there’s a page that has high traffic, but also high bounce rates. So for Scopa in this example is able to get visitors onto this page, but they’re not sticking around. They’re not interacting with the content. So that’s the auto identify piece. The auto suggest is providing a recommendation for what to do on that content to improve the engagement rate. You can see here that the recommendation is that the headline lacks a specific and differentiated benefit statement and the subheading is a bit long and might lose the reader’s attention. So the suggestion of course is to provide a more clear unique benefit statement and make the subheading more concise and impactful. So that’s the out of suggestion. The auto optimization is actually generating a variation of this experience based on those suggestions. So now we see a new heading here, the world’s finest coffee is delivered right to your door. Very specific benefit statement and a subheading that’s more concise, more clear, and more impactful. So from here what the market can do is go into their AM editor and tweak the content if the agent doesn’t get it quite the way that they want. And once they’re ready, they can come back into sites optimizer and they’ll have two options available to them. If they want to, they’re very confident that this new variation is going to move the needle in terms of engagement, they can select it and deploy and this will write the content as the new default experience for this page. If they want to run a simple A-B test, they can select multiple variations, click test multiple, and this will start an A-B test and will monitor the performance of these variations over time. And once a winner is detected to a statistically significant degree, they can come back in and promote the winning variation to the new default experience so that we can resolve this engagement.

There are other types of optimizations available in Site Optimizer as well. There’s optimizations for Site Map to improve search engine indexation. Also metadata issues are identified, suggested, and optimized to improve SEO and user experience. There’s also trained a model on security vulnerability detection and optimization. So we can see when there are issues like cross-site scripting vulnerabilities that are within the implementation and provide very prescriptive recommendations on how to resolve these types of vulnerabilities in the site projects. Other security vulnerabilities include CORS configuration issues, versioning issues that also user group permissions configurations. Continuing to add accessibility capabilities as well. So you see a full accessibility report based on your implementation of the accessibility rules down to WCAG 2.2 level A and AA standards and the specific issues that identifies and provides suggestion for as well. So this is a weekly report that you get from Sites Optimizer and will be archived as well. So you have a history of your accessibility stance against those standards over time. So that’s a quick tour of some of the capabilities currently within Sites Optimizer. The way that this is built, it’s kind of a framework or platform where we’re going to continue to add opportunities for optimization over time as well. So this will continue to grow. So with that, I’m going to hand it over to Meryl to take us through the development some of the optimizations and some of the learnings that we’ve gotten from customers. Thank you, Raymond. And just in time, I was finishing an answer to a question.

Let me share my screen, guys. Perfect. And actually just spot on because I was answering questions that made this input as well. The way we’ve been building Sites Optimizer over the last month or the last year is not coming out of the blue or out of what we think would make sense. Actually, we started building Sites Optimizer by doing manual work with customers in co-innovation mode. So everything that you have seen in Hyman’s demo and that is actually available in the actually available product has been actually productized after several manual interventions on customer side. And we keep working this way for every new functionality that we have. So Sites Optimizer is actually an ongoing and ever-growing product. I believe that we’ll always surface new opportunities from several types. What you’ve seen today is just a small sample of what we are actually currently working on and already enabling customers with.

So let’s see that as three steps. The first step is actually engaging with customers in co-innovation mode on their website, on their live website, on their live project and feeding the possible pain that they feel after one of the various success metrics that are there. Karthik was speaking about organic traffic acquisition, for example, or high bounce rates coming from traffic and being on the product page, for example. So we are actually helping customers do that. And the other learnings, whether one of the other techniques that we’re are actually leading to success. And once we are able to repeat this once, twice, three times, then it’s actually the right time to use the AI capabilities underneath to automate that scale.

The second step of the co-innovation is that actually when we are presenting one of the opportunities, we are aiming to provide the best flow in terms of user experience. There were questions on the chat around how does it work with my private pet repository? It’s not or how does it work in case I’m using this type of offering and not the other? Well, this is exactly another kind of precious feedback where we are iterating upon to provide the best UI in the product. Also the UX, we keep in mind that it’s a very user centric application. So it’s AI first, but the humans remain in power and in control at every stage. There is nothing that goes into production without having been reviewed and merged manually, of course. And the last part of the product validation is actually once we have the automation in place, whether this is very early with auto detection or mid-stage with auto suggestions and complete stage with auto optimize as well, customers who are working with us benefit from that at the super early stage and also can share feedback about the volume. Actually working in co-innovation for example, doing edge delivery projects with us. It’s nothing new for others. It might be something pretty new, but it’s not like we are taking the place of an implementation partner. We are working also in collaboration with many implementation partners in context of a customer project. It’s also not like we are shipping free POCs. It’s actually a commitment on both sides for our engineering team. It’s a commitment to actually make you successful as per your needs. But in return, we also expect the customers to be committed to work with us. It needs to be like ongoing daily in interaction. If we want a fast-paced interaction, which leads to a new functionality available in the product. We have plenty of customers in co-innovation with us right now. Here are a couple of logos from the very early adopters we had starting a couple of months ago, if not earlier for a few of them. Let me share with you a couple of success stories so that you can picture better how those co-innovation work. The first customer Wilson. If we have tennis fans in the chat or in the audience, you will know them. They are selling tennis related material, but also sportswear. Wilson have a storefront, so they sell products online. They have product data, product listing pages. They run a lot of marketing campaigns or paid campaigns to attract visitors and to actually aim to sell products that they want to promote. Yet using the operational telemetry, what we observed is that there was an issue with the cookie concern banner, which was actually just blocking the access to the whole page, but also not showing the product image, the main image of the product on the product page. So the bounce because people are landing on it, where you don’t see anything. So we’ve been helping them to run an experiment that was actually aiming to alter the look and feel of the cookie concern banner while preserving the legal requirements of it, of course. And we’ve learned that the challenger number two, which is the one showing completely the image on the product detail page, led to a huge business impact. So we’ve been repeating this a couple of times with other customers. And guess what? It’s a functionality that is getting productized as we speak. We have an agent underneath which is running exactly the same logic than we have been doing manually. It’s looking at the page, it’s taking screenshots of the page, it’s actually making recommendations on how the page would look like, also where the cookie concern banner should be positioned and helps to watch the environments for these challengers that can then be run through an experiment for validation. Another example, very simple, especially for the ones who are familiar with the performance work we’ve been doing over the last years, PGA tour, by the way, on HTTP delivery services with time, witnessed the entropy of the website that needs to live and to implement new use cases for new events that are, of course, very important for the brand. At the same time, page speed becomes maybe less of a concern. And using SiteOptimizer, we were able to show the customer that the CLS impact was huge, at least on desktop, and worked with them to actually fix the core vital degradation and bring it back to success because, of course, having a huge CLS, just like it was the case here, was degrading drastically the experience for people on desktop visiting the page and leading to higher bounce rate. So this example, among others, is actually guiding us now to improve the way we are auto-suggesting core vital degradations in the product in a very prescriptive manner. Because if you look at typical software out there, we are actually having generic recommendations most of the time, which are difficult to interpret for people who are not expert in the 2PIC. The next step is that we are working on generating the PRs as well. So this is not available today, echoing the question I had on the 2PIC as well, because we want to send PRs that will be meaningful and not just send PR for the sake of saying we send PRs, while actually it’s not going to help the person who has to review it and merge it at all. BemboHR, for this customer on the B2B space, it’s super important to be present on Google SERP and also on LLM responses to LLM users asking, of course, on specific topics for non-run it keywords, but also long-term keywords and others. And here, SiteOptimizer helped them a lot to actually run better for critical non-run it keywords, but also achieve a set of tasks that they were fully aware of but never had the capacity to take care of. So it’s not like a niche case, it’s very common that on the tech SEO side, especially or even on the accessibility side, customers know they have to do things, they have a backlog full of these kind of issues, they don’t have the capacity to do it. And in this case, they are empowered with a full-fledged capability from auto detecting to auto optimizing today, which leads them to be able to fix these kind of issues in one click, plus one review on the buffering side. Perchies is one of my favorites because you can see that we have been surfacing a lot of different topics with them. So it’s, of course, echoing the previous customer cases with page speed increase, SEO improvements, engagement improvements by running experiments. If you can see, for example, the screenshot on the slide, there was no CTA, especially not on mobile. So there was a paid campaign leading traffic to this page and there was no call to action. So traffic was bouncing. But so new topics like accessibility, on which we are working hard in this half to make actually the functionality available to all and security, where we also looking at how can we automate security fixes for customers because we know it’s the kind of issues that are potentially harmful towards reputation branding, but also capacity for internal fixes. I think that’s it on my side. So I would like to give some place and some room for the questions. Goran, I think there are still many questions pending.

Yes, thank you, Merl. And thank you also to Karthik. And hi, man. Yes, we have a couple of questions. Let us go through them. So the first one is could you I suggest to look at it because this is a long question. So could you clarify the license metric for sites optimizer, the product description page it says is content request. Is that CR for AM sites? How do the two correlate giving optimizer is separate license? And are there any generative action limits for how to suggest and how to optimize actions? I can take that one. So the license metric for sites optimizer is content requests. So it is similar to what we have with AM sites today. Now that it is a separate license. But the way we size for how many content requests you need for sites optimizer is similar to how we do it for AM sites. So today, if you have AM sites for 10 million page views, contract over 10 million pages, we will recommend that you’d have sites optimizer for 10 million content requests as well. And there’s a separate price point attached to that. There is no generative actions limit for what you saw the auto suggest an auto optimize.

Yeah, hopefully that answers the question.

Thanks Karthik. If not, please follow up with follow up questions. Okay, then the next question.

You’ve answered many already. Thank you for that. It does require a target license for A-B test. We are leveraging the AM contextual experimentation by default. If your customer has or your company or customer has A-B target, you’re fully free to use A-B target to run your test as per your preference. At this stage, we do not have an integration between the two products. This is definitely something that we are looking at.

And the next question, in case one site optimizer is enabled, does it introduce JS pages, JS to pages to understand content engagement on the site? So every AM cloud, every site running on one of the AM clouds, whether this is AMS, CS or HDG services is instrumented by default with the operational telemetry. So there is nothing that needs to be done at the project level to do so. With a small nitpick, if we would look at the headless website, this is not done because we have no control on the head and the front end code. So only for headless deployments, we would need the project team to actually inject the script in the in the petrology. Thank you. Can you talk about A-B test feature? Can you ask a more explicit question? I guess on a very high level.

Hey man, you want to take it? Yeah. As Mel mentioned, there’s what’s called contextual experimentation that’s available in AM was originally developed for edge delivery services. We’ve made it available across cloud service and AMS now as well, where you can run simple A-B tests, leveraging operational telemetry as the data source for, you know, determining winners in those A-B tests. So that’s the technology that it uses for the A-B tests that are started from sites optimizer. As Mel mentioned, we’ll be working with the target team and customers if they use target as their decisioning engine, but kind of out of the box by default, the A-B testing that’s part of sites optimizer, leverages the native capabilities that are built into AM now.

Thank you. I can take the question, go on about the potential extension points on the roadmap and the agentic approach. Sounds good.

There are the extension points on the roadmap. They are not where you may think they are right now. And I will also answer another question related to headless. We don’t think we’ll be able to auto-optimize for headless deployments because they’re heavily customized. And in that sense, we rather think that we will provide extension points for that, for project teams to actually contribute and be what’s required for these specific scenarios that are heavily customized and on which we will not be able to provide prescriptive recommendations and fixes. On the agentic side, we don’t have any integration with A-B capabilities as we speak today, as Ayman was also mentioning, but the architecture that we have in place is actually embracing open standards. So we are looking at MCP to plug the tools and bring them to agents and also A2A protocols. So I would rather advise you to look at this open standard to get started. And when we will open the gates for XMCP at this level, it’s likely the path that we will choose here.

Thank you, Marilyn. I think we have one more question. So please feel free to post your questions. And also in addition, I would like to ask you to rate this webinar as feedback is very much appreciated by you. You can also tell us what you would like to see covered in the future JAM session. So let me quickly post the link to the poll and to the chat.

There you go. It’s in the general chat. Okay. Now switching to the Q&A again. So the question was about the sandboxes. Just a second. Can it be provisioned for sandbox environments as well? If so, what needs to be done to get it? We are working on sandboxes now. They should be available soon. We’re starting with edge delivery services. So we can work on identifying, suggesting, and then auto optimizing into sandbox edge delivery services from a sandbox into an edge delivery services project. And then we’ll make the sandboxes available for cloud service following that. So more to come very soon. Not available just yet, but we are actively working on our sandbox strategy and approach, and we’ll share more very shortly. Thank you, Ayman. So we answered all the questions. We’re going to wait another moment to see if more questions appear. Otherwise, we will slowly close the session. There is one question about whether it’s available for customers who are using AM on prem. The answer is no. We are looking only at the cloud customers. Yeah. Thanks. I missed that one. Also for your information, when I will repost the contextual thread as well, you will find the recording and the slides there tomorrow or the day after. And I will post the link into the general chat in a second.

And there you can also ask a question post session. If there is something you really need to know and you can’t find that, then just please post your question there as well.

Post session. All right. There’s another question coming in. Is this GA now answered by Ayman? Yes.

Okay. We’re going to wait for another couple of seconds and then close the session if no questions are available. I really appreciate it if you have additional questions for us. We’re certainly available. Okay. Great. Then thanks a lot to the team for providing this presentation and insights and for answering questions. Thanks a lot to the audience for joining and for your attention and also for your questions. This is also important feedback for us. And also don’t forget to please rate this session and propose future topics for future webinar. And with that, we’re closing the session. Thank you very much. Have a great day or evening. Bye-bye. Thank you all. Thank you everyone.

recommendation-more-help
5f9e433e-d422-4bfd-9e43-c9417545dc43