Adobe Experience Platform

In ongoing modernization discussions, we are exploring how Adobe Experience Platform can be introduced gradually into existing Adobe Commerce ecosystems without requiring an immediate full migration.Key areas we are analyzing:event streaming strategy customer/profile data synchronization App Builder extensibility patterns coexistence of legacy customizations with composable services operational impact during phased adoptionWould like to understand from the community:What integration patterns are working well? Any lessons learned from real implementations? Common challenges around scalability, governance, or data consistency? How are teams preparing for future ACCS adoption while still running PaaS?Looking forward to hearing different implementation experiences and architectural approaches.

We are seeing more clients asking about progressive modernization instead of full replatforming.In several Adobe Commerce projects, the discussion is no longer only about storefront replacement, but also about how Adobe Experience Platform (AEP), App Builder, and event-driven architecture can coexist with existing PaaS implementations.Some common questions we are evaluating internally:How are teams introducing AEP capabilities without disrupting current Commerce operations? Are you using App Builder primarily for integrations, custom APIs, or operational workflows? What are the biggest architectural challenges during phased modernization? How are you handling legacy custom modules during transition planning? Are clients already preparing for ACCS readiness through composable services?Would love to hear how other teams are approaching:phased migration strategy event-driven integrations operational governance data synchronization patterns AEP adoption challengesInterested in learning fro

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)?

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

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

Hi Adobe Community, Greetings! For a certain use case, we were exploring options to create calculated fields from events source table and map it to a profile type schema. Will this work?In the events source dataset, there are fields which says what is the account type of the user, whether its active or not (Active flag), Identifier, etc.What we wanted to build in AEP was to create flag/boolean fields for a particular user, accounting for each account type. For ex - A user can have multiple accounts, say Savings account, Spending account, tax saving account, high returns account. What we want to create is 4 fields for each user. So the schema would essentially have Identifier + 4 account type fields.  Logic of each field-Identifier: would be mapped to Identifier. Savings Active: if Account type = Savings account AND Active account = True, then Savings Active field turns true else remains false.  Spending Active: if Account type = Spending account AND Active account = True, then Spending

We’re currently evaluating an enterprise API migration strategy for an AEM Cloud implementation and wanted to get input from teams that already went through something similar.Historically, our downstream AEM services relied on a middleware layer that handled:response transformation, field normalization, validation mappings, and compatibility logic for consuming components.Now that we’re moving closer to direct enterprise API consumption, we’re debating whether it’s better to:Introduce a transformation/adaptor layer temporarily to preserve compatibility for existing AEM services/componentsORRefactor all consuming services directly against the new API contractMain concerns are:scalability, long-term maintainability, regression testing effort, performance overhead, and dependency coordination across multiple teams.Would be interested in hearing how other enterprise AEM Cloud teams approached this tradeoff in production environments.

I am trying to add aria-label to rte link plugin. I have added it but its getting stripped even after adding the value to the xss protection config. How to solve this.

We have a non-XDM compatible dataset enabled for upsert with a couple of fields of type Object[].rootNode.fieldName: Object[]rootNode.anotherFieldName: Object[]We want to be able to edit one of them at a time and to replace the whole list on that update operation, including emptying out the list. On the dataflow mappings we have that field mapped asupsert_array_replace(rootNode.fieldName) -> rootNode.fieldNameWe were able to see full list replacements where the profile attribute had [A, B] and we sent [C] and it worked correctly becoming just [C].Our issue is that we couldn't clean the field from [A. B] to [] by sending [] or null, or anything else we could think of on the HTTP merge-enabled call.How can we clean an attribute of type list in that case?

We are exporting a CJA dataset from AEP to Azure Data Lake Storage Gen2 using the destinations connections in AEP. We have connected the account through the initial setup. However, we are receiving an error for the export job. We have confirmed the folder and path do exist and the account has all rights necessary to create the export. The documentation states: Folder path: Enter the path to the destination folder that will host the exported files.” We aren’t sure if this is an absolute path or relative path, but tried both. Any ideas?Error Code: ACTBTC-2400-400Error Message: An error has occurred when exporting data to the destination. Unable to perform the data export with the parameters flowId=45f4d5d5-62fb-4407-8c87-806a66e76015 datasetId=69a9f4aa23235bf67317dd6e flowRunId=550339e9-dded-431d-a115-47e4ef53afc5. The error message returned by the destination is An unsupported operation has been performed. The path may not be correctly specified.  

If the user is in two user groups:high level of access - edit  low level - view Which one has the priority? 

I had a question regarding reducing latency for data availability in Customer Journey Analytics (CJA).We understand that a latency of around 1–2 hours for data to become available in CJA is fairly normal. However, there is a requirement from the client side to have data available sooner, ideally within 15 minutes.One suggestion was to look into setting up a Streaming API to reduce the delay. From my understanding, the Edge Network to AEP/CJA setup already uses streaming ingestion, and based on Adobe’s SLA documentation, the processing time still appears to be within a similar timeframe.I’m aware that true real-time capabilities would require additional licensing/capabilities, which the client currently does not have. So my question is:With the existing AEP/CJA setup and licensing, are there any recommended approaches or best practices to reduce the latency and have data available sooner in CJA?Would appreciate any guidance from anyone who has worked on similar implementations.

We are trying to migrate our Snowflake authentication method to key/pair as recommended. We are following all necessary steps but still running into authentication error. Curious if anyone has done this lately and can share gotchas? 

Hi, I need to speak with an AEM cloud expert regarding a migration from on prem to cloud. Could you kindly get in touch with me. Thank you!. 

Hi Community, we are ingesting customer profile data into AEP via batch ingestion into a profile-enabled dataset. The batch runs show as successful in Monitoring and the data is visible in the dataset preview. However the Real-Time Customer Profile is not reflecting the updated attributes and our segment membership is not updating even after 24 hours. Identity stitching is configured with ECID as primary identity. We verified the dataset is profile-enabled and the schema has the correct union schema. Batch status shows as Success with no failed records. What could cause profile not to update even when batch ingestion succeeds?

A question that comes up frequently when working with shared Field Groups: does marking a field as Primary Identity in one schema affect other schemas using the same Field Group? And how do you handle required fields that don't apply to your specific dataset?These are two separate concerns that get conflated because they both live in the schema layer.Primary Identity is Schema-LocalMarking a field as Primary Identity is a local decision scoped entirely to that schema. It has zero upstream impact on the Field Group definition or any other schema using it.Identity designation and field group constraints are completely separate layers in AEP's Schema Registry:- Field Group = shared blueprint with structural rules (required, data type)- Schema = local implementation including identity role assignmentsField A stays required everywhere because that constraint is defined at the Field Group level. Field B being Primary Identity in Schema X doesn't change anything for other schemas.Handling Req

How is that possible? Or only Dev2 is on a prod?  

Hey team, I recently developed a data element to dynamically populate the datastreamID based on the page hostname - however the edgeConfigID is still resolving to %data element%. Any advice on how I can configure the data element so it is available at run time? Currently on Adobe Experience Platform Web SDK  v2.34.1 extension  In this thread it mentions that Adobe fixed the issue at the extension level , however I am encountering the same issue Thanks,Ty

Hi All, I’m looking to tidy up fields in our schemas that are activated for profile, and have been looking into field deprecation.  I fully understand deprecating a field doesn’t delete it, just removes it from the schema view, but I was also hoping it would clean it up from the profile view, audience builder…  Am I doing something wrong or is this by design and it will only be deprecated from the schema itself? Many Thanks, Paul

Hi, I need some help regarding an issue with our Adobe Experience Platform Web SDK and Adobe Target implementation.We are working on an integration using Adobe Launch, Web SDK (sendEvent) and Adobe Target, where everything works correctly on the initial page load. We already have a working setup where the Web SDK sends a request using correct decisionScopes, and a separate applyPropositions rule is used to render the experiences.However, we are facing an issue in a SPA environment where the page does not fully reload. Instead, filter changes update the URL using hash changes (window.location.hash).The problem is that when the user changes filters:the URL hash updates correctly but we cannot reliably trigger a new sendEvent the Launch History Change event is inconsistent sometimes it fires multiple times for the same filter change sometimes it does not fire at all depending on navigation flow As a result, the Web SDK request is not consistently sent, and in some cases we do not receive

What conditions need to be met in order to set up one to many relationship?It has to be experience event schema?