Adobe didn't ship new features on Monday — it shipped a new layer. CX Enterprise, built on MCP and A2A with every major AI platform partner signed on, is the real story from Day 1.
Shantanu Narayen walked out for his last Summit keynote as Adobe's CEO on Monday, and he brought Jensen Huang with him. The symbolism was deliberate — outgoing founder-CEO, arguably the most consequential chip executive of the decade, a Las Vegas ballroom full of Adobe practice leaders — but it wasn't the news. The news is that Adobe stopped talking about AI features inside Experience Cloud and started shipping an agentic layer on top of it, wired to open standards, with every foundation-model vendor that matters on the partner slide.
If you manage an Adobe practice, a digital experience org, or an enterprise IT budget that's been absorbing Experience Cloud spend for the last five years, Day 1 was the day the center of gravity moved. Not because any single product will transform your Tuesday. Because Adobe picked a lane — open agent standards, multi-LLM interop, human-in-the-loop orchestration — and committed to it publicly, with partners and GA dates attached.
This is the recap you need if you skipped the ballroom or sat through it jet-lagged. The through-line: agentic CX is leaving vendor-locked pilot purgatory, and the retooling bill is landing on your team.
The real headline: standards, not features
Adobe's load-bearing announcement was Adobe CX Enterprise, a new end-to-end agentic AI system that sits across the Adobe portfolio. Read that carefully. It is not a rebrand of Experience Cloud. Experience Cloud is still Experience Cloud. CX Enterprise is a new agentic layer — orchestration, governance, decisioning — that uses Experience Cloud as one of its substrates.
Two things make this more than marketing architecture theater.
First, CX Enterprise is built on Model Context Protocol (MCP) and Agent2Agent (A2A). These are the open interop standards the industry has been converging on over the past year. Adobe choosing them over a proprietary agent framework is the signal. It means your internal agents, your SI's agents, and Adobe's agents can — in theory — hand work to each other without a custom adapter per pair.
Second, the foundation-model partner list reads like the entire market: AWS, Anthropic, Google Cloud, IBM, Microsoft, NVIDIA, and OpenAI. All seven. Adobe didn't pick a horse. If you're a CMO whose board has asked why you're standardizing on one LLM vendor, CX Enterprise gives you an answer: you're not, and your DX stack doesn't have to either.
Anil Chakravarthy, Adobe's president of Digital Experience Business, framed it this way: “Adobe CX Enterprise enables businesses to scale agentic AI with a fully customizable solution that is tailored to the needs of their organization, moving teams beyond AI experiments to tangible business outcomes.” The key phrase is beyond AI experiments. That's the pitch and, honestly, the pressure every Adobe buyer is feeling from their CFO right now.
What ships now versus what ships later
This is the slide most enterprises will re-read on the flight home.
Shipping now: Marketing Agent in Microsoft 365 Copilot
The Marketing Agent is generally available in Microsoft 365 Copilot today. That's the one deployable item from Day 1. It's also in beta across Anthropic Claude Enterprise, ChatGPT Enterprise, Gemini Enterprise, IBM watsonx Orchestrate, and Amazon Quick — which is a reasonable preview of where agentic marketing work will actually happen: inside the LLM surface your employees already live in, not a new Adobe UI.
If your organization is on Microsoft 365, you can begin piloting this now. You do not need to wait for Coworker. Start there.
Shipping later: CX Enterprise Coworker
The headline product, CX Enterprise Coworker, is generally available in the coming months. No firm date. Coworker is the agent that orchestrates CX workflows end to end — synthesizing data, content, and decisioning across the fragmented systems every enterprise actually runs. It integrates NVIDIA OpenShell secure runtime and NVIDIA Nemotron open models, which is the posture for regulated industries that can't ship prompts to a public endpoint.
Anjul Bhambhri, SVP of Customer Experience Orchestration, called Coworker “the next step for businesses retooling marketing workflows to leverage agentic AI.” Retooling is the operative word. Coworker is not a drop-in; it's an implementation project, and the org design work should start now, not when GA hits.
The underrated story: Brand Visibility and the 269% jump
The data point from Monday that will age best: AI-driven traffic to U.S. retail sites rose 269% year over year in March 2026. That's new Adobe telemetry, and it's the number you cite when someone asks why you need a line item for AI discoverability.
Adobe's answer is the Brand Visibility solution, which ships with three agents: Brand Experience Agent, Content Advisor Agent, and Brand Governance Agent. Pair that with Adobe LLM Optimizer and AEM LLM Apps and you have the beginnings of an answer to the question every CMO got from their CEO in Q1: how does our brand show up inside ChatGPT and Gemini, and who owns it?
Today, nobody owns it in most orgs. SEO teams are adjacent, PR teams are adjacent, brand teams are adjacent, and none of them have the tooling or the mandate. Brand Visibility is Adobe trying to create the category before Salesforce does.
Product-by-product: what changed
GenStudio
GenStudio for Content Marketing launches as a new module. GenStudio for Performance Marketing now supports ChatGPT Ads — which is worth its own meeting with your paid media lead. If ChatGPT Ads becomes a real channel in 2026, your performance stack needs prompt ops and brand ops wired into media buying, not bolted on.
Firefly and production workflows
Adobe Firefly Creative Production for Enterprise Workflow Builder targets batch production — the unglamorous work of localizing 400 variants of a campaign. This is where creative ops budgets actually go, and automating it is a cleaner ROI story than most generative AI pitches.
AEM and conversational commerce
AEM LLM Apps, Adobe LLM Optimizer, and Adobe Brand Concierge together represent a real shift in what AEM is for. Classic AEM is a CMS. AEM plus LLM Optimizer plus Brand Concierge is an AI-discoverability and conversational-commerce surface. That changes what good looks like on an AEM team. Edge Delivery Services skills matter more. Classic component development matters less.
Real-Time CDP and Engagement Intelligence
Real-Time CDP expanded profiles now unify structured and unstructured data — transcripts, images, email corpora, the stuff that never made it into a segment before. Engagement Intelligence is a new decisioning engine optimized for customer lifetime value. This is the backend most buyers will under-invest in and regret.
Journey Optimizer
Journey Optimizer Loyalty adds a module for gamified, loyalty-status-aware journeys. With agent-driven journey edits in the CX Enterprise vision, AJO architects with governance chops become scarce fast.
Also announced
Adobe Brand Intelligence (continuously-learning engine for on-brand content), a 3D Digital Twin Solution with NVIDIA, and a Workfront Workflow Optimization Agent. The partner list — Accenture, Capgemini, Cognizant, Deloitte Digital, EY, IBM, Infosys, PwC, TCS on the SI side; Dentsu, Havas, Omnicom, Publicis, Stagwell, WPP on the agency side; Adyen, PayPal, Stripe on payments; Acxiom, Demandbase, RainFocus, SAP on data; 24/7.ai, Algolia, Netomi, Genesys on CX — is comprehensive enough that CX Enterprise will have real implementation capacity at launch. Jon Gieselman, Xfinity's CGO, came on stage as a confirmed customer, which is worth more than another demo.
What this means for your team
The honest part. Every product above implies roles most Adobe practices don't staff today.
Agent ops is a new function. Someone has to configure skills, author governance policies, and run the human-in-the-loop review that keeps an agent from approving the wrong discount. This isn't a martech admin role and it isn't a data science role. It's closer to an MLOps-plus-brand-governance hybrid, and the people who will be good at it are currently scattered across your RevOps, marketing ops, and platform engineering orgs.
AEM teams need to split. Classic AEM developers are still necessary — the CMS doesn't go away — but the growth headcount is in AEM plus AI discoverability, with Edge Delivery Services fluency. If your AEM shop can't answer a question about LLM Optimizer today, that's the gap.
Performance marketing needs hybrid talent. Paid media buyers who also do prompt and brand ops. ChatGPT Ads is not a channel you run with a 2024 playbook.
Data engineering shifts toward unstructured. Real-Time CDP ingesting transcripts, images, and email means engineers who can model unstructured corpora — not just SQL and identity graphs.
AJO architects need governance experience. Loyalty journeys edited by agents raise the bar on approval workflows and audit trails. That's a governance skill, and it's rarer than it should be.
None of this is a 2027 problem. The Marketing Agent is GA now. Your team either has the roles or it doesn't.
Where this leaves buyers
Adobe made the right bet on Monday. Open standards beat proprietary lock-in for enterprise buyers in a multi-LLM world, and the seven-partner model list gives CIOs cover. The execution risk isn't the technology — it's organizational. Most enterprises will buy CX Enterprise in 2026 and under-deliver on it in 2027 because the roles to run it don't exist on their org chart yet.
That's the execution gap we talk about at Focus GTS. If you're thinking through what agentic CX means for your Adobe practice — the hiring plan, the Navigator support model, or just a sanity check on which roles to post first — get in touch or email us at info@focusgts.com.