Adobe Summit 2024 - what Workfront’s evolution tells us about Adobe’s composable journey

Phil Wainewright Profile picture for user pwainewright March 29, 2024
The quest for end-to-end automation, accelerated by AI, is leading enterprise applications towards a composable architecture - Workfront's evolution provides a glimpse of how that's going at Adobe.

Anil Chakravarthy stands on stage at Adobe Summit 2024 in front of a colorful graphic of the Content Supply Chain
Anil Chakravarthy on stage (Adobe Summit 2024 screengrab)

Scrolling through the session catalog on the opening day of Adobe Summit this week, it was notable that most if not all the Workfront sessions were already full. The huge interest in the work management tool, which Adobe acquired in 2020, comes as marketing teams wrestle with the accelerating pace of content creation and campaign delivery that’s now possible thanks to the increasing automation enabled by generative AI. They need the end-to-end project management control that Workfront gives them — and which Adobe is increasingly embedding in other products that make up what it calls the content supply chain.

This process of embedding Workfront functionality into other products is symptomatic of a broader phenomenon across many enterprise application vendors, all driven by the same need to ease end-to-end automation for their customers. The old landscape of discrete products that only connected in limited ways is being superseded by a highly interconnected panorama of diverse components. While each individual product continues to evolve, it’s doing so in ways that allow each of its component functions to be available elsewhere. Chris O'Neal, Principal Product Marketing Manager at Adobe and a Workfront veteran, explains:

All the different Adobe applications, as we build to the future, are looking to architect those future capabilities as composable services, so that they can be utilized where it's appropriate in other venues outside of the core product that it was built for.

The apotheosis of this approach can be seen in the all-new Adobe GenStudio, which brings together a whole range of functions to create, edit, deliver and evaluate the performance of content and campaigns created with the aid of generative AI. He goes on:

When you build a new app like that — a net-new app in the Adobe universe — it can use or consume those composable services that are bringing those functions. The person on the other end doesn't really know that, 'Oh, this piece came from Workfront, or this piece came from Adobe Express or this piece came from [wherever].' It's just seamlessly integrated into that UI.

Rethinking existing functions

While it's straightforward to create new functions as services, such as the newly introduced Firefly Services which provide generative and creative cloud models and functions as APIs, this new approach often means rethinking how functions have been built into existing applications. For example, Workfront and, Adobe's media collaboration app, have both had content review and approval functionality. Those are now being converged into a single shared function, taking the best from each. O’Neal says: has some great capabilities around specifically content review. What's going to be happening over time is that we're bringing in that technology around the content review. The approvals engine will stay within Workfront and will be driven by Workfront. So we're unifying the review and approval platforms.

When it comes to the end-to-end content process, there’s been a rethink of how assets are managed within that process. Previously, digital asset management (DAM) was thought of as one step in the process, but now Adobe is positioning its DAM platform, AEM Assets, as a service that is available at every step. The content supply chain then becomes more of a cycle or ecosystem than a linear process. Don Bennion, Head of Adobe GenStudio Product Marketing, says:

What we're trying to do is illustrate this less as a completely linear waterfall process, but more as an end-to-end, cyclical process that our customers are going through to consistently and constantly optimize your content supply chain.

Infusing generative AI into this process places new demands on each application to provide context around the data it uses, which can then frame instructions or prompts that direct the AI to provide accurate, reliable responses. One of the enhancements that Adobe has been working on for Workfront is the development of a marketing operations data model and data graph. This underpins an important new planning capability that has been launched this week, but will also provide a framework that an AI assistant can use to interpret and import unstructured data from documents such as campaign plans. This is a time-saving alternative to the traditional method of inputting the information manually, or preparing it in another system and then importing the data. O'Neal says:

A lot of planning still happens in documents. There's a campaign brief, or a project outline in Word, or Powerpoint, or a PDF. The AI assistant, you can literally drag and drop a document into the planning area, and it will ingest that and then read the content, pull out the different things that you want to track out of it. Of course, you're going to adapt it or reconfigure some of those things as it may not be perfect, but it's an automated way of doing that, so you don't have to start from scratch.

Introducing Workfront Planning

The introduction of a new planning module is a significant upgrade to Workfront, as O'Neal explains:

The Workfront that most people know and love today, a lot of that has to do with workflow, a lot of that is execution-based, when you already know what you want to do, and you're making the assignments, building out the projects, aligning resources. But there's the upstream aspect, I need to model out what that whole operation looks like first, before I know what projects to even create or who to assign to this.

With Workfront Planning, it really is the end-to-end operational view. Let's say you want to plan campaigns for the coming year. You can literally take it from, what the overall goals of the campaign are, all the way through its definition, and then all the way through all the different tactics around audiences and channels and targets, and also attaching budgets to it. So when we say planning, we mean it in a holistic way. It's everything from the goal all the way up to the customer touch points and back.

The underlying marketing operations data model and data graph give organizations the flexibility to define their own custom fields and the relationships between them, and then create views that allow them to see where resources are needed and how projects have performed. He goes on:

Not only can you define your own custom records, you can have them relate to each other in any hierarchy that you need them to. So you can define exactly what your operational workflow is, exactly as you need it. And then off of that data set, we also produce curated views — do you need a marketing campaign calendar, click this button, here's your calendar. Do you need a spreadsheet style view to see all the data associated with his campaign? Click this button, here's that view.

The evolution of a more componentized architecture also means that functions and data from Workfront and other Adobe applications will even be accessible from within third-party applications. For example, a new partnership with Microsoft announced this week will make it possible for workers in Microsoft 365 apps such as Outlook, Teams and Word to get answers and insights via the Copilot AI assistant from Adobe applications such as Workfront and Customer Journey Analytics, and incorporate the information into documents they're working on.

My take

The latest enterprise applications are being built in a highly composable Tierless Architecture that makes data and functions available as needed. This gives established vendors the challenge of refactoring their existing products to co-exist with this new paradigm. Workfront's journey is a case in point. Its functions must become available for embedding in other applications, both within the Adobe family such as GenStudio, or even beyond as in the case of the Copilot for Microsoft 365 example. In some cases, that means adopting functions from companion applications in place of existing capabilities, such as the integration of review functions from, while is adopting Workfront's approval functions. The advent of generative AI is not only accelerating this trend towards more composability, it also influences the make-up of the new architecture, as we see in the definition of graph databases that map the entities and relationships that each application deals with, providing more context around the data to help AI make sense of it.

It's evident from this glimpse of what's happening at Workfront that Adobe is taking this transition seriously and is making significant strides in the right direction. But there's still much to do — even before thinking about the potential to build Workfront capabilities into Document Cloud, a topic on which Adobe has remained tight-lipped although I sense we may hear more on that front before long. For any vendor of Adobe's stature, these are huge changes to undertake. But they are essential to support the end-to-end automation and AI capabilities that are becoming table stakes for today's enterprise applications.

A grey colored placeholder image