AEM Assets and Dynamic Media video updates in 2026.9.0

The 2026.9.0 feature release of Adobe Experience Manager as a Cloud Service, dated 24 September 2026, puts most of its weight on Assets and Dynamic Media. Content Hub gets AI Search, Smart Collections and custom metadata sorting, and Dynamic Media video gets a new customizable viewer, AI-generated captions and attribute-based access control. If you run a digital asset library or publish video, this is the part of the release to read first.

The release also carries platform items such as canary deployments and Java timelines, and plenty of coverage will lead with those. This post stays with the asset and video side, because that is where marketing and content operations teams will feel the change.

Key takeaways

• The current feature release is 2026.9.0, released on 24 September 2026. The next, 2026.10.0, is planned for 29 October 2026.
• Content Hub adds AI Search, Smart Collections and custom metadata sorting. They reward libraries with clean metadata and punish messy ones.
• Dynamic Media video gets a customizable viewer with actionable analytics, AI-generated captions, multi-caption and multi-audio support, and attribute-based access control.
• Multi-caption and multi-audio tracks, and the Dynamic Media Template component in Adobe Journey Optimizer, sit in the Limited Availability section. Pilot them, don't promise them.
• Audit metadata, captions and access rules before enabling anything. The new features amplify what is already in the library.

What changed for Assets and Dynamic Media in 2026.9.0

The release overview groups the Assets work into two themes. The first is discovery and organization in Content Hub. The second is richer Dynamic Media experiences. Here is the list as the notes frame it, with a plain reading of what each item is for.

Notice what is absent from that table: anything that forces a migration. These are capabilities you adopt at your own pace. That shapes how to plan, and it is why the question that matters is which of them solves a problem you really have.

Who should read this, and who can skip it

Three groups should act.

The first is the DAM owner. If your teams complain that they can't find approved assets, or that the library is a graveyard of near-duplicates, the Content Hub changes are aimed at you. Intent-based search and automatically maintained collections both attack the same complaint from different sides.

The second is whoever owns video. If you ship product, training or campaign video in more than one language, the caption and audio-track work is directly relevant. It matters even more if accessibility requirements are pushing you to caption everything.

The third is the person who owns rights and permissions. Attribute-based access control for secure delivery is a governance feature. It deserves a conversation with security or legal before anyone configures it.

Who can skim? An AEM Sites team that only uses Assets as a back-office store, with no Content Hub, no Dynamic Media and no video, will find little here that changes their week. They can read the platform items instead and come back later.

Content Hub: AI Search, Smart Collections and metadata sorting

AI Search is described as intent-based and multilingual. In practice that means a marketer can type what they are trying to do or show, rather than guessing the filename or the exact tag someone used two years ago. A regional team searching in their own language and still reaching the right global asset is the obvious win.

Be realistic about the dependency, though. Search that understands intent still works from what the library holds. If half your assets have no descriptions, inconsistent tags and cryptic filenames, better search will help less than you hope. It will surface the assets that are well described and bury the rest, which can be worse than a keyword search that at least returns something predictable.

Smart Collections stay up to date automatically. The notes don't spell out the rule mechanics, so check the documentation for how membership is defined before you design around it. The use case is clear enough, though. Think of the collection somebody rebuilds every quarter: approved assets for a product line, current logos for partners, imagery cleared for a region. If a rule can describe it, a Smart Collection can keep it current, and a person stops being the point of failure.

Custom metadata sorting is the quiet one. It sounds minor and it is the item most likely to change daily behavior. If reviewers can sort by expiry date, campaign or rights status instead of only by modified date, a lot of manual checking goes away. That only works if those fields are populated, which brings us back to the audit.

A sensible order for a DAM team:

1. Export a sample of assets and check how many have a meaningful title, description and the metadata fields you rely on.
2. Fix the schema first. Decide which fields are mandatory and which are optional.
3. Collect twenty to thirty real search phrases from different teams and languages.
4. Run those through AI Search and note where results disappoint.
5. Build two or three Smart Collections for jobs people currently do by hand.
6. Add sorting on the one or two metadata fields reviewers actually ask for.

Don't announce a relaunch to the whole company at step one. Let a pilot group tell you where it breaks.

Dynamic Media video: the new viewer, captions and audio tracks

The release describes a new, fully customizable video viewer with actionable analytics. For teams on an older player, the draw is control over look and behavior plus reporting that someone can act on. The notes don't say which metrics are included, so don't assume parity with whatever analytics tool you use today. Compare the two side by side on a real page.

AI-generated video captions are the feature with the clearest operational payoff. Captioning a back catalog by hand is slow and expensive, and many teams simply haven't done it. Generated captions give you a draft. A draft is not a finished accessibility asset. For anything public, regulated or technical, have a person review the output, particularly product names, acronyms and speaker names. Sample twenty videos across your content types before you trust the rest.

Then there is multi-caption and multi-audio track support. The Limited Availability section describes Dynamic Media with OpenAPI capabilities now supporting multiple captions and multiple audio tracks, associating several language-specific tracks with a single primary video, managed from a unified interface. The point is easy to see. Instead of publishing five copies of a video for five markets, you keep one primary asset with the tracks attached.

Two cautions apply. This is under the Limited Availability heading, which is Adobe's signal that it is not yet generally available and may change. And the notes describe the capability, not the production workflow around it: who writes translations, who approves them, how a missing track is handled. Those questions are yours to answer.

Attribute-based access control and secure delivery

Attribute-based access control for Dynamic Media is the item that most needs a decision before any configuration. The idea of attribute-based control is that access depends on properties, such as the type of content or the audience, instead of a hand-maintained list of people. The release notes give it only a short mention, so read Adobe's documentation for exactly which attributes and delivery paths are supported.

What to do before configuring:

• List the video and image assets that must not be publicly reachable: embargoed launches, internal training, partner-only material.
• Write down, in plain language, who should see each group and why.
• Check which attributes already exist in your metadata and which would need to be created.
• Test a delivery URL as an unauthorized user. Confirm the asset really is blocked, not just hidden from the page.

That last test is the one teams skip. A viewer that doesn't show a video is not the same as a delivery URL that refuses to serve it.

Limited Availability and beta: how to treat them

The notes are clear that LA and Beta features stay listed until they become generally available. Two Assets items are in the LA section: the multi-caption and multi-audio support, and the native Dynamic Media Template component in Adobe Journey Optimizer.

The second one lets marketers add and personalize Dynamic Media templates directly in the Journey Optimizer email editor. It covers selecting existing templates or importing layered PSD files, editing layers such as countdown timers, and configuring text, calls to action, positioning and timing. The notes mention open-time countdowns and personalized product content as examples. If your email team and your DAM team are separate, this one needs both in the room.

The beta programs are a different thing. The notes include a strong disclaimer: beta releases can contain defects, come without warranty, and features and APIs can change without notice. They list programs for agents in AEM, AEM Foundation, Cloud Manager and AEM Assets.

A rule that keeps teams out of trouble: never put an LA or beta feature on the critical path of a launch. Use it in a sandbox or a low-stakes campaign, and keep the old way of working available until the feature is generally available.

The rest of the release, briefly

The overview also covers things outside Assets. AEM agentic capabilities are now generally available in Adobe CX Enterprise Coworker, the AEM MCP Server adds Cloud Manager operations, and IDE agent skills expand with AEM query optimization and additional code-migration patterns. On the platform side there are canary deployments, OpenTelemetry-based APM (Limited Availability) and IMS authentication rich errors.

There are also plan-ahead notices on Java API deprecations, with a date of 28 September 2026, and the upcoming Java 25 runtime upgrade. Developers and architects should read those sections in full, because custom code is where deprecations bite. For a DAM or video team, the useful move is simply to ask your developers whether they have looked.

How to test and measure the Assets changes

Do this in order, and write the results down.

1. Baseline first. Before enabling anything, record how long it takes a marketer to find a known asset, how many support requests the DAM team gets about finding things, and what share of videos have captions.
2. Pilot with one team. Pick a group that searches a lot and tolerates rough edges.
3. Compare search outcomes. For your list of real phrases, record whether the right asset appeared on the first screen, before and after.
4. Review a caption sample. Score accuracy on the content types you care about, including names and jargon.
5. Test access rules from the outside. Use an unauthenticated browser session and a user without the right attributes.
6. Check the viewer on real pages. Look at load behavior, accessibility with a keyboard and screen reader, and what the analytics actually report.

For measurement, avoid inventing a target before you have a baseline. Direction is enough at first: faster time to find, fewer duplicate uploads, more captioned video, fewer manual collection rebuilds. If the numbers don't move after a month, the cause is usually metadata, not the feature.

What not to do

Don't switch on AI features across the whole organization on release day. Don't treat generated captions as final for anything that carries legal, medical or safety weight. Don't plan a multilingual video program on a Limited Availability feature without a fallback. And don't skip the metadata cleanup because the new search sounds clever. Intent matching can't recover facts that were never recorded.

Also resist the urge to chase every item in the release. The next feature release is planned for 29 October 2026, and there will be another list after that. Pick the two changes that remove a real pain, adopt those well, and let the rest wait.

Where a partner helps, and where your team can manage alone

Your internal team can handle a lot of this. Auditing metadata, defining Smart Collections, choosing sort fields, sampling caption quality and running a small search pilot are DAM administration jobs, and a capable admin can do them with Adobe's documentation.

A partner earns their fee in a few specific places. One is a multilingual video workflow that spans translation vendors, approvals and delivery. Another is access-control design when many audiences, brands or regions are involved. The third is integration work, such as connecting Dynamic Media templates to Journey Optimizer. A partner also helps when your schema needs a redesign, because that touches every team and every integration that reads metadata.

If you do bring one in, write the brief around outcomes: find time, caption coverage, protected delivery. Our RFP template for Adobe Experience Manager gives you a structure for that, and you can browse Adobe Experience Manager partners to see who works on Assets and Dynamic Media specifically. If you are still working out which kind of firm fits, find partners lets you filter by need. For more release coverage, the blog has related posts.

A useful test for any partner pitch: ask them to show how they would measure success on one of the changes above, using your baseline. Vague answers there usually predict vague delivery.

Sources

• Adobe Experience League: Current Release Notes for Adobe Experience Manager as a Cloud Service — Release dates, 2026.9.0 overview, beta programs, and the Assets and Dynamic Media sections.