Coming out of Adobe Summit 2026 this year, there is an undeniable and overarching theme: Agentic AI interactions and behavior are the new way of how content is developed, created, and optimized. For organizations using Adobe Experience Manager (AEM) and Edge Delivery Services, the implications for how we author content, migrate experiences, and develop code are massive.
Here are some of my top learnings and takeaways, which I fully expected but were firmly confirmed after attending the sessions, labs, and co-hosting my own Ask Me Anything session on the topic of Agentic AI.
1. The Rise of Customer Experience Orchestration: Adobe CX Enterprise & CX Coworker
One of the most significant announcements at Summit was the shift from traditional marketing to full-lifecycle Customer Experience Orchestration (CXO) with the introduction of Adobe CX Enterprise. This agentic system is designed to unify the content supply chain, customer data, and brand visibility into a cohesive platform optimized for both humans and AI agents.
A feature within this ecosystem that I’m particularly interested in is the CX Enterprise Coworker. It’s more than just an automated chatbot. It sits alongside the authoring experience, serving as a persistent “AI teammate.” You provide the business goal, like launching a campaign, and the Coworker builds and executes the workflow behind the scenes. It continuously monitors performance signals, adjusts actions in real time, and orchestrates specialized agents (such as Audience or Journey Agents) to keep projects moving efficiently without constant manual oversight.
2. Conversational Content Authoring & Migration with DA.live
Content operations are getting easier and quicker with natural language processing. With the new Document Authoring (DA) MCP Server, AEM has introduced conversational content authoring in DA.live. The Model Context Protocol (MCP) provides tools and prompts that allow AI agents to interact directly with AEM services, which will drastically simplify agentic content creation and management.
Instead of manually clicking through a traditional AEM interface, authors can use chat interfaces to ask the system to perform complex tasks, like finding all pages missing a specific component or updating campaign banners across a microsite.
In addition, a new interface “Experience Workspace” provides a new canvas to author and preview content, while a browser extension is also being developed to assist in migrating content to EDS-friendly component blocks and styles.
For content migration, Adobe showcased the Experience Modernization Agent. This tool automates the migration of legacy websites to cloud-ready formats, helping teams onboard to Edge Delivery Services through a simple chat interface that iterates on updates in real-time.
3. Keynote Spotlight: The New Rules of Brand Visibility
If there was an “ah-ha” moment for attendees at the conference this year, then this was it. The day two keynote, “The Elevated Brand Experience: Orchestrating Discoverability for the Agentic Web,” focused on what we should all be hyper-focused on now: AI agents are increasingly browsing catalogs and evaluating specifications on behalf of human buyers. This means that traditional SEO is no longer enough; brands must now optimize for Generative Engine Optimization (GEO). I’ve been highlighting this more as we ask customers about their GEO strategy, and Adobe stressed this even further.
Because AI hallucinates when it encounters unstructured data, websites must now feature “Self-Describing Content.” This was also a main point in Tom Cranstoun’s recently published book, “MX: The Handbook.” The keynote emphasized that building “Computational Trust” with AI models requires pristine underlying code structures and semantic metadata. To help enterprises manage this, Adobe highlighted the LLM Optimizer, a tool that automatically identifies where content is invisible to AI and deploys optimizations so your brand remains cited, relevant, and prominent in agent-driven search results.
4. Vibe Coding for Edge Delivery Services (Lab L613)
Probably one of my favorite sessions and targeted to developers, the hands-on lab “L613: Vibe Code an Author Experience for Edge Delivery Services” was an absolute highlight. The session demonstrated that developing with AI coding agents (like Cursor, which we used in the Lab) for Edge Delivery Services requires proper orchestration so agents don’t make assumptions about your project’s structure.
To make agentic development super-productive, developers can inject context into their projects using AGENTS.md and llms.txt files, which feed AI agents proper project conventions and AEM best practices. The lab also showcased the use of specialized “Skills,” such as the content-driven-development skill, which acts as a mini-playbook forcing the AI to create test content before writing block code, ensuring the AI agent correctly models the content, implements CSS, and reviews the output against established patterns. We built some experiences in less than an hour, which I would have estimated to take days in traditional development cycles.
Looking Ahead After Adobe Summit
Adobe Summit 2026 made it clear that we are moving from AI assistants to true orchestrators. Focusing on these new agentic workflows, conversational authoring interfaces, and GEO strategies, we fully expect brands to scale their content supply chains and secure their discoverability in the agentic web. The agentic web isn’t just coming—it’s already here. It’s time we start building for it. If you’re wondering how to prepare your AEM architecture for the agentic web, or want to discuss your GEO strategy, let’s connect.
