Product Communities

As generative AI and Retrieval-Augmented Generation (RAG) continue gaining traction across enterprise platforms, I’ve been thinking about how current AEM content architectures will need to evolve to support future AI use cases. Many AEM implementations were originally designed around page rendering, content reuse, and headless delivery. However, AI-driven experiences introduce additional requirements around content retrieval, context quality, metadata strategy, and knowledge organization. Some areas we’re currently evaluating include: • Content Fragment design and granularity• Metadata quality and governance• Taxonomy structures for improved retrieval accuracy• Structured vs unstructured content strategies• Content duplication and source-of-truth management• GraphQL delivery patterns for AI-powered applications• Preparing enterprise content repositories for future RAG architectures One challenge I see is that AI systems are only as effective as the content they’re able to retrieve and

Are you exploring Adobe Edge Delivery Services (EDS) for your next AEM project?Want to understand:• What EDS actually is• Why Adobe moved from Helix → Franklin → EDS• How EDS differs from traditional AEM Sites• Why Lighthouse performance scores matter• How the edge architecture works in real-world projects  Click here to read the complete guide:https://www.aemcq5tutorials.com/tutorials/what-is-adobe-edge-delivery-services-eds/

Hi team,I’m working on a 6sense + Adobe Launch integration and facing an issue with how data appears in Adobe Analytics reporting.Current ImplementationOn a first‑time visit, the page loads normally and fires a page view call, followed by a non‑page view call that contains the 6sense attributes (stored in eVars). I’m trying to avoid delaying the page‑view call.If anyone knows how to include 6sense attributes in the initial page‑view call without delaying it, please suggest. On a return visit, I merge the 6sense eVars into the page‑view call.ProblemIn reporting, I see unexpected behavior.Example: In the Country dimension, United States and Unspecified both show 70%+ for Visits—together exceeding 100%.(Refer to attached screenshot.)Any best practices to include 6sense attributes in the initial page‑view hit without delaying it?

Check out the upcoming user group meetings happening in June/early July:☀️ [Virtual] Wednesday, June 17 @ 6pm ET: Conversando con Pablo Larrosa sobre la evolución de Adobe Experience Manager y el ecosistema Martech (Spanish-speaking event)​ with the AEM - LATAM Chapter☀️ [Virtual] Thursday, June 18 @ 4pm CET: Bank Security, Fintech Speed: A Modern Bank Website on AEM EDS Document Authoring​ with the AEM – DACH region chapter​☀️ [Virtual] Thursday, July 2 @ 4pm CET: Building an AI-Powered Tagging Engine on AEM Assets​ with the AEM – DACH region chapter​ Sign up here: https://aem-augs.adobe.com/events/#/list

Join the Adobe Champion Program and become a part of a global community of passionate product experts who are driving thought leadership and innovation!  Adobe Champion Applications are open through June 5.The Adobe Champion Program offers an opportunity for growth, connection, and influence. As a Champion, you’ll join a distinguished group of professionals who are not only advancing their own careers but also helping shape the future of Adobe’s most impactful products through shared expertise and community leadership.Each year Adobe selects 40 Champions across a number of products. These are:Adobe Analytics Adobe Experience Platform Adobe Commerce Adobe Workfront Adobe Marketo Adobe Experience ManagerWho Should Apply?Champions come from a variety of backgrounds, each with a unique set of experiences and insights into our products. However, there are a number of characteristics each Champion has: expert experience, passion, and a desire to engage with Adobe and fellow practitioners.Exp

I’m looking for more details about the custom action button that can be created in Content Hub. I recall there was some information shared about adding a custom action button to the asset card. I’ve attached a screenshot showing where these buttons would appear. Does anyone have more information on this? I checked Experience League for resources but wasn’t able to find much. 

I see a OOTB response header for EDS pages - Vary: Accept-Encoding,X-Forwarded-Host, is there a way to edit the values present in this header?

Hello, is there any possiblility to make the comment when confirming a “Assign Task” step in the AEM inbox mandatory?Currently it is only optional on all cases but we need the possibility to also be able to force users to add a comment. Thank you! 

How are you all documenting your analytics requirements to pass to developers? We release new features/products on our website and app monthly, and are currently using the same drawn-out and tedious process to request our analytics implementation work that we’ve been using for years. We (the analysts using Adobe Analytics) receive designs of the experience in Figma from Product, screenshot those screen by screen and outline what variables/events need to set on each screen/with each interaction in a PPT, and then a technical specifications document gets created by someone on the more technical side (using trackState, etc.) in a PDF. That document is then passed to IT for development. There has to be a more efficient way. I’d love to hear how all of you are handling this process when you’re pushing multiple analytics implementations across multiple platforms monthly.

Hey! Looking forward to hearing more about this! My core ideas to test will include: Agentic Marketing Development Representative Agentic Content Recommendation Engine Agentic Audience Profiling Agentic Personalized Content Agentic Social Signal Listener I’ve been working on what the ideal integration looks like as well as what the core primary Resources should be, aka which MD files should be contextually prefixed to the prompts. What are people doing for: Skills Repository and Sharing Cheers,Edward UnthankFounder, Etumos 

Hi there,I’m exploring how to connect Adobe Experience Platform (AEP) with Copilot Studio and/or Claude to build agents and reusable skills on top of AEP data.Do we currently have any MCP server, native connector, or prebuilt integration available for AEP in this context?If not, could you share any recommended guidelines or best practices for creating skills using AEP data (e.g., via APIs, custom connectors, or middleware)?

In AEM Authoring (especially AEMaaCS), when working with Tab or Accordion components, there is a noticeable UX issue where after editing a component in a non-default tab/accordion panel, the view resets back to the first tab upon closing the dialog. This creates friction for authors when working across multiple tabs or deeply nested components, as they are required to manually navigate back to the previous context after each edit. Current Impact:Slows down authoring productivity Causes confusion, especially for business users Impacts usability for complex pages with multiple tabs/accordions Requires repetitive navigation, which is inefficient at scaleProposed Enhancement:Preserve the active tab/accordion state after dialog close Maintain scroll position and UI context during authoring Optionally persist author state (session-based) for better editing continuity  Why this matters:With increasing adoption of structured components (Tabs, Accordions, Containers) in AEM Cloud implementation

I would like to understand CJA in depth so here is my question. Additionally I understand SQL well (I assume CJA is SQL  on the back-end….)Assume I have two events with the same structure - XDM paths: Event1 - add to cart  {   product.price: 55,      eventType: addToCart   }   Event2 - purchase {   product.price: 33,      eventType: purchase   }   How to create metric which tracks only product.price when eventType is purchase?

Adobe Commerce performance mistakes I see in almost every auditAfter dozens of Adobe Commerce (Magento) audits across B2C and B2B storefronts, the same culprits keep appearing. And they're not exotic , they're avoidable.MISTAKE #1 : VARNISH MISCONFIGURATIONCache hit ratio under 60% on category/PLP pages.What I find: ESI not set up properly, TTL values too conservative, cache invalidation firing on every CMS save.Fix: Audit your Varnish VCL, separate CMS cache from catalog cache, use hole-punching for cart/account blocks.MISTAKE #2 : SYNCHRONOUS INDEXING IN PRODUCTIONFull reindex triggered on product save, blocking storefront performance.What I find: indexers still set to "Update on Save" instead of "Update by Schedule".Fix: Switch all indexers to schedule mode. Use cron for execution. Use partial indexing wherever supported.MISTAKE #3 : N+1 QUERIES IN CUSTOM MODULESCustom GraphQL resolvers hitting the database once per product in a listing.What I find: Third-party or custom modules byp

Our migration experience has been challenging to say the least. Our assigned techs have tried their best to get everyone on the same page but we’ve had gaps in getting back to us over the past few months. We recently were told all users of the cloud have to backup their data before we migrate. We asked for directions to send out to our users and Adobe does not have any we are told. Support keeps kicking us back to our techs for the answer and our techs tells us to contact support...you get the idea. We’re 11 business days until Adobe will ‘force migrate’ us regardless if we’re ready. Any idea what happens when that happens? Trying to be positive but we’re on a three day streak of non-response to questions. I want to provide our almost 7K end users the best heads up on what is needed to be done in advance. Christina

Recently some of our documents attached to projects are giving an AWS error when opened.  This seems to be happening with issues that are converted to projects with fusion (convert to project module).  We have Preserve Issue noted so the document attaches automatically to the new project.  The attachment is fine on the request but corrupt on the project. I have a support ticket in but haven’t been able to resolve.  Anyone else running into this issue?

Hi Team,I have a Velocity token created in Marketo that you previously helped with to display a dollar amount, round it off, and calculate it as 10% of the revenue. The token works perfectly when added within the body of the email.#set( $tenPercent = $number.integer( $math.mul( $lead.personLevelEngagementScore, 0.1 ) ) )${number.format("${esc.d}${esc.h}${esc.h},${esc.h}${esc.h}${esc.h}", $tenPercent)}When I wanted to display the same value in the subject line, adding a $ before the token worked successfully. However, I now need to use the same token in the preheader, and neither adding just the token nor adding a $ before the Velocity token seems to work there.I’m not sure if there’s a limitation with preheaders or if there’s another approach required to make this work. Could you please help guide me on this?Regards,Akshat

Has anyone been experiencing an issue where a (Proof HQ) stage deadline is removed from an automated proof after it has started? We have seen it sporadically over the past couple of months, but is hard to reproduce for me to place a ticket.Any insight into a cause or a possible solution would be appreciated.Thanks,

Hi Community, Do we know what will happen when classification rule build gets deprecated? our team is looking to get some insights on what we should do to prepare for this as most of our classification is being populated using rule builders. Thank you in advance!

我目前使用的是Substance Painter 11.0.2 / 12.0.2，遇到了一个关于复杂蒙版的重大工作流程问题：当图层的遮罩图变得复杂（包含多个生成器、层级、绘画和滤波节点）时，手动合并这些遮罩节点常会导致软件崩溃，使过程非常不稳定。 我正在寻找一个能一键自动完成以下操作的插件：保留原始图层（完全可编辑，非破坏性，便于回滚） 将最终遮罩结果烘烤/平整成一个干净的黑白纹理贴图 创建一个新的空填充层，并自动将烘焙的遮罩纹理分配到遮罩通道 最终结果会是一个没有遮罩子节点的新图层，只使用单个纹理作为遮罩。这不仅提升了性能，也消除了处理复杂遮罩图时发生崩溃的风险。虽然手动变通方法（重复→组→平整组）有效，但非常重复且耗时。如果有一个简单的插件来自动化这个流程，解决常见的掩膜相关崩溃问题并简化工作流程，那将非常有帮助。

We have a custom text block used across 100+ pages in AEM Edge Delivery Services.Our block structure is based on ordered JSON fields, and recently we introduced a new field in the block. Because of this, all existing row indexes shifted by one, and we updated our JS/decorate() logic accordingly.However, after deployment: The updated JS expects the new field order. Existing pages still render plain.html in the old structure/order. The new field is missing in the rendered output. modelFields/children count still reflects the old structure. If we manually edit and re-save a page/block (even a minor change), the block gets regenerated correctly: modelFields gets updated children count becomes correct new field appears in the front So currently it looks like all existing pages would need manual re-authoring/re-saving.One additional complication in our setup: We cannot safely add new fields at the end either. Our tabs are structured like overview + common. The common tab append