Adobe Summit 2024 - Adobe harnesses generative AI to automate end-to-end CX processes

Phil Wainewright Profile picture for user pwainewright March 26, 2024
Summary:
At this year's Adobe Summit, new product announcements focus on using generative AI to automate end-to-end customer experience processes for marketing teams.

Adobe Summit 2024 stage @philww
(@philww)

At last year's Adobe Summit, generative AI brought new capabilities to creative design with the launch of Adobe Firefly. This year, the technology is bringing new levels of automation to the entire process of content creation and delivery in ways that will redefine how marketers go about their work. As the opening keynote kicks off this year's event, Adobe is announcing a series of innovations in how brands create, deliver and measure customer experiences.

With the emphasis on end-to-end processes, today's announcements are far more interlinked than is usually the case for this type of annual big vendor event. While there are individual products in the mix, the big theme is how they all work together, with a massive sprinking of generative AI providing the glue that connects them. Pride of place goes to Gen Studio, first unveiled last October, a new product which acts as a control plane across the existing Adobe portfolio to co-ordinate the creation, deployment and measurement of creative assets and campaigns.

A new AI assistant for the Adobe Experience Platform provides a conversational interface so that users can more easily access capabilities within the platform. Other new announcements help automate the procees by which brands can create and deploy personalized marketing. It's all about taking generative AI and providing a toolset that allows brands to work faster and more accurately to co-ordinate Customer Experience (CX) data, content and journeys. In a pre-briefing ahead of today's news, Amit Ahuja, SVP of the Digital Experience Business at Adobe, explains:

The majority of the brands that we work with are still largely trying to figure out, candidly, what to do with Gen AI. It's in an experimental phase, both at the practitioner level, even at the corporate level, in terms of what they can use, what they should use of those use cases. I think Adobe has an amazing opportunity... How do we bring Gen AI in a very responsible way to marketers to go from an experiment to actually help them make their jobs easier and better?

An AI assistant for Experience Platform

In the Adobe Experience Platform AI Assistant, Generative AI comes to Adobe's own applications with a natural language interface. Adobe has fed knowledge of its own products into LLMs to produce a provides a simple conversational interface that can answer technical questions and will automate tasks, simulate outcomes and generate new audiences and journeys within Adobe Experience Cloud applications. The product uses customers own data to frame prompts so that the underlying LLM is able to generate an accurate answer. Users can ask questions to uncover high-performing customer segments for instance, or ask the AI Assistant to create a segment that is more likely to embrace a certain promotional offer. Ahuja explains:

How do you inject an assistant type of technology directly into the surface area of our platform and our applications that enables you to effectively ask questions, automate tasks, and start creating and enabling a lot of things like, how do I create the right audience? How do I create the right content, all these different pieces. We think there's a massive opportunity to help our customers where they are in the application ...

This [assistant] understands the context of the user, and, most importantly, understands the context of the enterprise data. So when you're saying, 'Hey, I want to understand a segment,' or 'Hey, create a segment for me of the people that are highest propensity to book a hotel next week,' as an example, the thing that we're really dong is classifying all the enterprise data, just with their data ... We can use that natural language interface to help them answer those questions much faster and better.

Personalization at scale

Other new features support the goal of providing easier, better personalization at scale, along with the ability to more easily try out different customer journeys and analyze performance feedback at scale, even down to individual attributes. Ahuja comments:

One of the major things that all these customers are asking for is, hey, in the context of personalization at scale, help me experiment faster, better and easier, and across different things. I don't want to just experiment across different types of content. I want to experiment across different types of audiences. I want to experiment what is the right journey ...

Not only do they want to be able to experiment faster, they want more people to run more experiments — they're saying, 'Hey, I have a whole team of people. How do I make this really easy for them to experiment, but using the right set of data and using the right set of controls?' So this specific announcement is we're announcing a lot more unified experimentation capabilities to be able to allow them to do that.

As an example, a new variant generation feature in AEM Sites allows brands to take a single marketing asset and create numerous variations of copy that speak to different audiences. For example, teams can generate variants of a web page which adapt the copy for target personas across different industries, job roles, age groups, and other parameters. New Adobe Content Analytics makes it possible to understand the performance of AI-generated content down to the level of attributes such as colors, objects and styles, analyzing which resonate best with target audiences.

A new unified experimentation capability enables brands to increase the volume of tests to pinpoint customer paths that maximize conversion and drive reuse of offers across channels. Brands can support experimentation across content, channels and audiences, with machine learning models, and centralized administration to push winners and resolve conflicts.

Adobe Journey Optimizer, in which brands orchestrate customer journeys, can now bring together brand-initiated and customer-initiated interactions within a journey, ensuring that as brands interact with consumers on channels such as the web, social media and mobile, any real-time changes in user behavior or preferences are captured, triggering new experiences.

Meanwhile, a new Adobe Journey Optimizer B2B Edition, which integrates with the Adobe CEP and its Marketo B2B marketing platform, enables sales and marketing teams to target specific buying groups within customer accounts, a departure from broad-based marketing. It leverages unified data in Adobe Experience Platform and generative AI to personalize the journey, through a web chat bot for instance.

Gen Studio, a 'GenAI-first' approach to content

Finally, Gen Studio brings together a set of capabilities as a template for how to create, deploy and measure marketing assets in the generative AI era. Now in preview and expected to become generally available in the middle of the year, it helps brands manage the end-to-end content supply chain, with direct connections into Adobe Experience Cloud and Creative Cloud applications including Photoshop, Express and Firefly, AEM Assets and Customer Journey Analytics, and integrating the project management capabilities of Adobe Workfront. Ahuja sums up:

We now have a GenAI-first approach to everything around content, but we bring it together into a new singular application ...

This isn't a standalone tool. We have so much rich technology across Adobe, what we're doing is really bringing that into this new service with a lot of new capabilities, but also allowing the power of choice wherever the user wants to go [if they want to go deeper].

This is in response to growing demand for faster creation, editing and feedback around what Adobe calls the content supply chain which is core to the customer experience in campaigns. He goes on:

In years past, the burden on the content supply chain maybe wasn't as high as it is now. The desire for experimentation, the desire for personalization, all of that is putting exponential pressure on top of this to deliver more and more iterations of content. And by the way, you put GenAI out on top of that, you;re able to iterate much faster, create much faster. I think that puts a lot more pressure on what is the business process across every part of the content supply chain.

It aims to provide a single, easy-to-use interface for the creation of on-brand enterprise-ready content, including the ability to use a custom AI model trained with proprietary brand assets to easily create on-style assets. Analytics are surfaced within the context of asset management and the content supply chain - for example, customer journey analytics. The product embeds prompt engineering to help ensure content conforms to brand guidelines, including a visual alert that shows a green check for content that's on-brand, or a yellow check if it strays out of brand. The generative AI within the platform will make suggestions, such as suggest different ideas for a call to action in a website. Ahuja says:

What is the intersection of the different types of content with different types of audience? You can go in and slice and dice and do all the types of visual and query analysis you want to be able to do there...

Is that content performing against those audience types? And then how do I continue to refine that, and what is the learning from that, such that when I'm going back to create more content, I can be much more on point to be able to produce the right content for that audience?

This is where I think we are very uniquely driving this full cycle across all these pillars and helping our different customers.

The full feature set of Gen Studio includes:

  • Workflow and planning - track and view campaigns, manage campaign briefs, see tasks assigned.
  • Creation and production - find images and generate variations through Adobe Firefly, including Custom Models to ensure that content is on-brand, unlocking personalization at scale for content creation and production.
  • Asset management new content hub to easily find assets, connecting to AEM Assets with the ability to create personalized variations with Firefly in Adobe Express.
  • Delivery and activation - create different variations of marketing assets for different distribution channels. Users are alerted if content is not on-brand, with suggestions on how to remedy it.
  • Insights and reporting - a feedback loop helps brands understand which generated assets and campaigns are performing best, down to the attribute level and attributes can then be used to inform generative AI prompts. Connection to Adobe Customer Journey Analytics drives a holistic measurement of customer experiences across channels such as web and mobile.
Loading
A grey colored placeholder image