Adobe CX Enterprise Coworker Hands Assistants the Front Door
Image design by Aragon, rendered by Canva AI.
By Adam Pease
Adobe CX Enterprise Coworker Hands Assistants the Front Door
On November 4, 2026, Adobe put its Adobe CX Enterprise Coworker agent inside the AI Assistants that employees already use. The company says the agent is now available in ChatGPT (as a plugin), Claude, Microsoft Copilot, Gemini for business and Databricks Genie. Amazon Quick, IBM and Snowflake are listed as coming soon.
The move puts Adobe’s customer experience orchestration agent where marketers already work, instead of asking them to come back to an Adobe application. This blog overviews the Adobe CX Enterprise Coworker announcement and offers our analysis.
Why Did Adobe Bring CX Enterprise Coworker into Third-Party Assistants?
Adobe describes CX Enterprise Coworker as an agentic teammate for customer experience orchestration. The agent reached general availability in June 2026. It is built on Model Context Protocol (MCP) and Agent2Agent (A2A), and it runs workflows on Adobe CX Enterprise with human oversight. Adobe says it handles campaign performance analysis, audience forecasting, journey optimization, conflict resolution, and planning and work management.
The announcement has two tiers. Adobe says five integrations are live now: ChatGPT, Claude, Microsoft Copilot, Gemini for business and Databricks Genie. Amazon Quick, IBM and Snowflake are coming soon, and Adobe gave no timeline for them. Adobe also launched pre-built MCP connectors to partner systems, including Amazon Ads, Demandbase, an ads MCP server for Meta campaigns, Microsoft Work IQ, RainFocus and Snowflake Cortex AI. The company says marketers can connect these systems without custom integration work and act across several of them with a single request.
Adobe explains the motive directly. “Today, marketers often use AI tools in isolation, without the organizational context and domain intelligence needed to personalize every stage of the customer lifecycle,” said Anjul Bhambhri, SVP of Engineering, Customer Experience Orchestration at Adobe. Adobe also named nine global systems integrators and six agency holding companies building on the agent, including Accenture, Deloitte Digital, Publicis and WPP.
Analysis
At Aragon Research, we view this as a clear signal that business applications now ship into the assistant rather than pulling users back into the app. Adobe made the right call. An agent that only runs inside its vendor’s interface gets used only when someone opens that interface, and many enterprises have already standardized on one or two assistants. Showing up in all of them keeps Adobe in the daily workflow.
The cost is real, though. When a marketer starts a request in ChatGPT or Copilot, the assistant owns the conversation, the history and the choice of which connector to call. Adobe keeps the data, the journey logic and the governance behind the request, but it gives up the daily user relationship. In Copilot, Adobe also sits inside a host that sells its own marketing and customer insights applications.
The problem extends beyond Adobe. The same open protocols that let Adobe plug into every assistant let every competitor plug in too. Over time, that weakens every app vendor’s hold on the customer, because the assistant becomes the place where work starts and where tools get chosen. App vendors will compete on the quality of what sits behind the request, not on the screen in front of it. Vendors whose data and workflow layers are only comparable will lose leverage to the assistant platforms.
Governance is the open issue. Adobe has not disclosed pricing, how it audits actions taken from third-party assistants, or whether features match across the five platforms. An agent that changes a live customer journey from Claude or Gemini needs the same approvals and audit trail it has inside Adobe’s own console.
What Should Enterprises Do About Adobe CX Enterprise Coworker?
Enterprises should decide which assistant is their front door. If business applications are going to ship into ChatGPT, Claude, Copilot and Gemini, letting each team pick its own assistant will scatter agents, data access and audit trails across platforms. IT leaders should name a primary assistant, and a secondary one if needed, and set the integration standard around it.
Next, govern agents across that front door, not app by app. Security and architecture teams should inventory which AI Agents can reach which systems through MCP connectors and who approves actions that touch customers. For Adobe customers, the pilot should start in a contained workflow, such as campaign analysis, before any agent changes a live journey from a third-party assistant.
Finally, ask Adobe hard questions before signing. Get answers on pricing for agent work that starts outside Adobe, on human approval and audit when actions start in another assistant, on feature parity across the five live platforms, and on delivery dates for Amazon Quick, IBM and Snowflake.
Bottom Line
Adobe CX Enterprise Coworker shows that business applications now meet users inside their AI Assistants. That keeps Adobe relevant but hands the daily user relationship to the assistant platforms, a shift that will squeeze every app vendor. Enterprises should choose their assistant front door now and govern agents across it before every vendor ships its own agent into it.