Adobe Research at Summit 2024: AI assistants, instant marketing playbooks, gen-AI infographics, and more

April 12, 2024

Tags: Adobe Summit Sneaks, AI & Machine Learning

Innovations from Adobe Research took center stage at this year’s Adobe Summit, the world’s largest digital experience conference, including exciting glimpses into the future of digital marketing and groundbreaking new features that are available now.

Unveiling the new AI Assistant in Adobe Experience Platform

During the Summit opening keynote, Anjul Bhambhri, SVP, Experience Cloud Platform Engineering (DX), announced beta plans for the new AI Assistant in Adobe Experience Platform (AEP) — a technology that was developed with major contributions from Adobe Research. The AI Assistant is designed to provide guidance for marketers, offering insights on everything from data and analytics to customer journey simulations. In the keynote, Bhambhri outlined four big things ahead for the AI Assistant: answering user questions, automating tasks, simulating workflows, and generating audiences and journeys. The AI Assistant was also a hot topic in the Summit strategy keynote, “Embrace Generative AI to Transform Experiences,” in demos at the AI Assistant’s booth, and across several labs and breakout sessions.

Adobe Research led the development of the AI Assistant, including the architecture, middleware, annotation, evaluation, impact quantification, and core models to create a productive and trustworthy user experience. The technology is designed to accurately detect a user’s intent, retrieve relevant document snippets, and harness the power of a large language model (LLM) to generate a response. Before a response from the LLM is presented to a user, it moves through a series of steps, developed by Adobe Research, to detect and remove hallucinations and attribute sources for the information. These steps build in accuracy and trust, so users can feel confident about the answers from their AI Assistant. Adobe Research also created a pipeline of high quality, real-world, annotated training data to continually improve the AI Assistant, including use-case-driven evaluation metrics and training and evaluation sets crowd-sourced from experts.

2024 Sneaks from Adobe Research

During Sneaks—the event’s annual preview of technology-in-the-works—Adobe Researchers presented AI-powered tools for creating instant marketing playbooks and powerful infographics.

Project Perfect Plays

It can take weeks to create and execute a marketing plan. But with Project Perfect Plays, powered by Adobe Firefly generative AI, teams can accelerate the creation of marketing playbooks and initiate marketing campaigns in Adobe Experience Platform. Users simply input an overview of their campaign objectives (or upload a marketing brief) and Project Perfect Plays generates a complete marketing playbook with audience insights, marketing strategies, and goals. And the whole plan can be kickstarted with the click of a button.

Collaborators from Adobe Research: Eunyee Koh and Chang Xiao (presenter)

Project Infograph It

Infographics help brands communicate complex ideas in easily digestible, interactive, and visually appealing ways. But the process for creating infographics can be technically difficult and time consuming. Project Infograph It simplifies the creation process, leveraging Adobe Firefly generative AI to create and recommend design elements based on key messages, data, and the overall aesthetic a user hopes to convey. Users simply import a dataset, enter a short text description of the top messages, and experiment with different color palettes, chart types, graphics, and animations.

Collaborators from Adobe Research: Gromit Chan, Eunyee Koh, and Tongyu Zhou (presenter)

Firefly-Powered tools in Creative Cloud for Enterprise

Summit’s opening keynote highlighted exciting new Firefly-powered tools for Creative Cloud for Enterprise. The tools, developed with Adobe Research, will help Adobe customers keep up with skyrocketing demand for new content that wows customers and drives business. Object Composites allows users to easily composite products into new scenes and swap different products in, all while keeping a consistent creative direction. Custom Models allows users to customize Firefly’s generative AI to their own brand style, products, and campaign direction.

Even more Adobe Research at Summit 2024

Adobe Research innovations also made appearances across Summit’s keynotes, demos, labs, and breakout sessions.

  • Object Completion is a new feature that uses generative AI to automatically complete hidden people and objects in an image, making it possible to move an element in an image to reveal what’s behind it. Presenters Shaquille O’Neal and Eric Matisoff wowed audiences with a peek at the technology before this year’s Sneaks.
  • Unified Experimentation in Adobe Experience Platform (AEP) and Adobe Journey Optimizer (AJO) was another of the big innovations announced at Summit. The technology allows marketers, product managers, and developers to progressively enable experimentation across content, channels, audiences, journeys, and machine learning models, all with centralized admin to push winners, resolve conflicts, and manage global holdouts. The technology was featured during the Summit strategy keynote, “Harness Your Data to Drive Personalization at Any Scale.”
  • Real-Time Customer Data Platform Collaboration helps solve the challenge of discovering, reaching, and measuring high-value customers without the use of third-party cookies. The technology allows advanced data collaboration between advertisers and publishers in the evolving, privacy-conscious marketing landscape. Real-Time CDP was featured during the Summit strategy keynote, “Harness Your Data to Drive Personalization at Any Scale.”
  • Token optimization for GenStudio is a new feature that reduces the token count of an input prompt without affecting the quality of the generations. The technology reduces costs and latency.
  • Asset Identity is a new behind-the-scenes feature in the insights tab of GenStudio. Asset Identity, which groups identical image or video assets, is implemented as a custom version of Content Authenticity Initiative (CAI) fingerprinting with strict matching of visual content.

You can find even more details about Adobe Summit 2024 here. And if you’re wondering what else is happening inside Adobe Research, check out our latest news.

Related Posts