When more than 14,000 marketers gathered in Las Vegas this April for Adobe Summit, they got a glimpse of how agentic AI will help them orchestrate the customer experiences of the future. Adobe Research was behind some of the most exciting announcements, sessions, and sneak previews at the event, from the unveiling of new tools that help marketers understand their customers’ behavior and design customer experiences at scale to hands-on agentic AI sessions. During the Sneaks, Adobe Researchers took the stage with demos of cutting-edge, in-development technology for designing with AI collaborators. 

Here are a few of the most exciting product announcements, sessions, and Sneaks that Adobe Research helped bring to this year’s Summit.  

Adobe Research powered big keynote announcements at Summit 

From the Summit main stage, attendees learned about game-changing new tools built on work from Adobe Research.  

Proactive Insights and Root Cause Analysis (RCA) are new agentic capabilities that identify the most important changes in business performance and explain why they happened. These tools fill a big gap: understanding critical business changes usually requires skilled analysts working for days, and the findings don’t always reach the marketing and business teams who need them. So Proactive Insights continuously monitors the full analytical space of a business to automatically surface verified, relevant, actionable insights such as drops in conversion rates or spikes in orders. Root Cause Analysis investigates the findings and provides marketers and business teams with verified explanations checked against live data.  

Adobe Researchers contributed key elements of Proactive Insights and Root Cause Analysis, including the intelligence that powers the analysis and scoring systems to determine which information gets surfaced and why. Researchers were also behind the evaluation benchmarks for continuous improvement, and a framework that personalizes results for each user.  

Adobe Research also built key capabilities for Brand Concierge, the intelligent AI assistant that helps marketers understand user intent and deliver the right content at the right time. Most recently, Adobe Researchers helped make the Brand Concierge experience more intelligent by continuously adapting to complex user requirements. Researchers also helped deliver an automated configuration experience and a personalization engine that enables Brand Concierge to deliver tailored, conversational customer experiences.  

Attendees met new products that understand their brands and orchestrate customer experiences 

Adobe announced Adobe Brand Intelligence, an agentic system that continuously learns about a business’s brand. By building a deep understanding of a brand’s preferences and patterns, Adobe Brand Intelligence makes it possible to apply an enterprise’s brand across the entire content supply chain, especially as AI systems become part of the workflow. Adobe Research contributed intelligent logo detection and variant matching, which ensure correct and compliant logo use. Researchers also developed the product’s ability to quickly and automatically analyze assets against brand guidelines and visually trace any violations to specific rules with multimodal attribution.  

Attendees were introduced to Adobe CX Enterprise Coworker, an AI-powered teammate that handles coordination and operational complexity behind the scenes, freeing marketers to apply their expertise and deliver experiences at scale. Marketers can choose an outcome, such as launching a campaign or responding to a market signal. From there, CX Enterprise Coworker builds and executes the workflow needed, and then adapts as conditions change. Adobe Research helped build the tool’s memory and context management capabilities, which allow it to handle tasks across complex, multi-step marketing workflows. Researchers also developed the tool’s as well as multi-agent evaluations, planning and orchestration capabilities, and optimizing skills. 

Marketers dug into the details and got hands-on practice with agentic AI tools 

Adobe Researchers helped shape Summit sessions where attendees learned about orchestrating customer experiences with Adobe’s agentic AI tools.  

Two complementary sessions, Audience Agent at Work: A Practitioner’s Guide to Smarter Workflows and Using AI to Build Audiences in Adobe Real-Time CDP, explored how marketers can create and manage audiences with AI. Adobe Research contributed the capability to create segments with natural language and the ability to create segments by combining previous user-created segments in Adobe Experience Platform.  

In the Unlock Insights Faster with Customer Journey Analytics Data Insight Agent session, attendees learned how conversational intelligence is revolutionizing exploration, visualization, and understanding without the need for manual queries or dashboards. Adobe Researchers contributed technology behind the Data Insights Agent and the capability to dig deeper into “why” questions and reveal hidden drivers behind patterns such as traffic spikes and revenue changes.  

Sneaks gave a peek at ground-breaking in-development customer experience tools 

Each year at Summit, Sneaks give attendees an exciting peek at in-development technologies that may shape the future of Adobe’s products. This year, Adobe Researchers joined comedian, actress, writer, and producer Iliza Shlesinger and Adobe Principal Evangelist Eric Matisoff on stage to share their work. 

Project Concurrent 

Research Scientist Tongyu Zhou presented Project Concurrent, a tool that tackles the problem of static infographics in a changing world. With Project Concurrent, users can bind live data directly to an infographic to automatically update changing information, like weather forecasts, concert tour dates, or daily stock prices. Users decide how often the system updates the visual properties of the design, ensuring that graphics stay current, eliminating the need for manual updates.  

Project Test Kitchen 

Research Scientist/Engineer Intern Yuzhe You shared Project Test Kitchen, an AI agent-supported design tool that gives users more ways to control image generation and allows teams with lots of ideas to collaborate. Users can enter a text prompt, reference images, and their own sketches to guide a design. From there, Project Test Kitchen offers options that vary along user-defined axes, such as composition and art style. After the first generation of ideas, users can extract favorite elements from a generated image and combine them in new ways, focusing and refining as they go. Project Test Kitchen replaces rigid text prompts with more space for ambiguity and an experience that’s more like the real-world design process.  

Project Face Off 

Research Scientist Doga Dogan presented Project Face Off, a tool that uses AI-generated personas to simulate expensive, slow A/B tests in just seconds. Users choose several possible website designs and then Project Face Off simulates how different audience segments would interact to each design, predicting which one is most likely to achieve the user’s specified goals. Users also get actionable insights, including the underlying reasons behind each persona’s preferences. This rapid “pre-flight” evaluation lowers the risk when marketers are choosing designs and can even help users explore ideas for niche audiences before launching expensive live experiments. This Sneak gained a special honor: It was voted the audience’s favorite.

Did you miss something? Catch up on everything that happened at Adobe Summit 2026 here.  

Wondering what else is happening in Adobe Research? Check out our latest news here.