top of page
A robot and a person build a plan on a table in front of a path of pieces leading to an arch with the sun

Adobe CX Enterprise Coworker: from a journey that thinks to a goal that gets executed

2 hours ago
3 min read

At the end of my previous article about Adobe Journey Optimizer in 2026, I asked: will your journey know when to stay quiet? Now comes the next question. What happens when a team stops drawing the journey, gives an agent a goal and reviews the plan it proposes?


That is what Adobe describes with CX Enterprise Coworker. But first, a note about availability, because the story has changed since April.


At Adobe Summit in April 2026, Adobe said Coworker would become generally available in the following months. On June 10, 2026, Adobe announced its general availability. That is Adobe's statement. I could not confirm the countries where it is offered, the required licenses or the actual price: the announcement mentions introductory pricing without numbers. Ask Adobe to clarify these three points before evaluating it.


From drawing a journey to defining the goal


According to the April announcement, Coworker is designed to work toward defined business goals by coordinating multiple agents and turning clear objectives into multistep actions. Adobe's example is a marketing team seeking a 3% improvement in cross-selling performance. Coworker can coordinate the relevant agents and tools to prepare a targeted offer, audience segments, creative assets and performance insights. It then drafts a plan and, once approved, helps execute the campaign and track performance against the goal.


Notice the sequence: goal, plan, approval, execution, measurement. Before, the team assembled the pieces and the journey was the plan. Now the agent writes the plan first, and the human work begins by reading it.


The June announcement adds that Coworker inherits existing data policies, consent rules and permissions from an organization's workflows. This is Adobe's claim; it is worth validating in your own environment rather than taking it for granted.


What changes with Adobe CX Enterprise Coworker in a marketing team's daily work


A marketing team reviews the plan that the Adobe CX Enterprise Coworker robot presents on a board

If Adobe's description works as intended, three things change:


  • Less manual assembly. According to Adobe, a natural-language instruction can lead an agent to build the campaign plan, content and journey flow for a defined audience within one process.

  • More review. The day becomes less about building and more about examining plans. Is this audience appropriate? Does the message sound like our brand? Is this the right channel?

  • Measurement closer to execution. Adobe says Coworker monitors performance signals and adjusts flows against defined goals, with integrated analytics to assess impact.


This is my interpretation, not Adobe's: the bottleneck shifts. It is no longer simply production capacity; it becomes the quality of the objective and of the review. An ambiguous goal produces an ambiguous plan, potentially executed very quickly.


What remains a human decision: when to proceed and when not to


Adobe's design includes an approval step. In April, it described greater automation while maintaining human oversight and control. In June, Adobe stated that teams approve plans, refine strategies and launch campaigns. Approval is built in; whether it is exercised responsibly depends on your team, not the product.


When it makes sense:


  • When the objective is measurable and clear, such as increasing cross-selling in a segment with reliable history.

  • When your data, consent rules and exclusions are already in order.

  • When someone with sound judgment will review the entire plan before approving it.


When it does not make sense. At least not yet:


  • When you cannot explain what counts as conversion or who should be excluded. The agent does not define those meanings for you.

  • When you would approve the plan without reading it because ‘the system already checked it.’

  • When no one knows who is responsible if an integration fails or the agent proposes something off-brand.


These are the same four questions from the journey, in a different disguise: which event moves the customer forward, which conditions exclude them, how will we know value was created, and who is accountable when something fails?


Defining the right objective, deciding what not to do and owning the outcome remain human responsibilities. I used to ask whether the journey would know when to stay quiet. Now I ask: will your team know how to reject a beautifully presented plan?


Sources:


Comments


​

  • Youtube
  • X
  • LinkedIn

©2025 by Victor Garnica.

bottom of page