---
title: CX Coworker
slug: hol-bayport/cx-coworker
createdAt: 2026-08-11T17:43:35.449Z
---

## Use agentic workflows

Below are use cases and sample prompts that practitioners are using in Adobe CX Enterprise Coworker Chat. Try these sample prompts to gain insight on the audiences built today

| **Use Case**                           | **Description**                                                                                                                   | **Skill(s)**                | **Application**       | **Sample Prompts**                                                                                                               |
| -------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------- | --------------------------- | --------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| Create audiences from natural language | Orchestrate step-by-step audience creation with user approval at each phase                                                       | audience-creation-flow      | Real-Time CDP (RTCDP) | “Create an audience of users who purchased in the last 30 days” · “Build a segment for high-value loyalty members in California” |
| Build PQL definitions                  | Assemble audience definitions from XDM properties, behavioral events, or existing audiences; support aggregation and time windows | segment-definition-assembly | Real-Time CDP (RTCDP) | “Create a PQL for people who viewed 3+ products but didn’t purchase” · “Add a 7-day time window to my event condition”           |
| Search & find audiences                | Find audiences by ID, name, semantic search; detect duplicates and analyze overlap                                                | audience-search             | Real-Time CDP (RTCDP) | “Find all loyalty audiences” · “Is there a duplicate of my ‘Holiday Shoppers’ segment?”                                          |
| Estimate audience size                 | Estimate profile reach for a PQL expression using Adobe Experience Platform Preview API with polling                              | audience-size-estimate      | Real-Time CDP (RTCDP) | “How large is this audience?” · “Estimate reach for this PQL expression”                                                         |
| Audience size waterfall                | Decompose a PQL into sub-predicates and show how each condition contributes to final audience size                                | audience-size-waterfall     | Real-Time CDP (RTCDP) | “Show me the waterfall for this PQL” · “Break down how each condition reduces the audience”                                      |
| Discover XDM fields for targeting      | Search fields by name, description, or data value; see where they live and where they’re already used                             | field-discovery             | Real-Time CDP (RTCDP) | “Which fields can I use to target loyalty customers?” · “Find fields related to purchase history”                                |
| Publish / save audiences               | Persist audience definitions to Experience Platform Segmentation Service with naming conventions and compliance checks            | audience-publish            | Real-Time CDP (RTCDP) | “Save this as a draft” · “Publish the audience with name ‘Spring Sale Buyers’”                                                   |





:::::Tabs
::::Tab{title="Use the Agent"}
## Create an Audience with Adobe Agent

In this exercise, we will use **Audience Agent&#x20;**&#x74;o create the audience.

Open the AI chat in full screen, and enter the following prompt:&#x20;

:::BlockQuote
<font color="#BE185D">Build an audience o</font>f visitors who add a product to cart then land on the checkout page but do not land on order-confirmation.
:::

![](https://api.archbee.com/api/optimize/1Jp0iolWv35Z_mkI7oUiT/o_6ugSFTzrSLMKLZfj0cd_screenshot-2026-08-06-at-21952-pm.png)

The agent will analyze you request and might search for an existing audience that matches the criteria. **Request to proceed with building it out anyway.**

It might then perform field discovery. In CX Coworker, you have the option of setting Business Context and a Data Dictionary. Providing this information beforehand allows the agent to learn in advance and will minimize Q\&A during the conversational chat experience.

Continue reviewing the audience definition and the size estimate. Proceed to create the audience. You will be asked if the agent should proceed with publishing the audience. **Proceed**.&#x20;

Verify by viewing your newly created audience

![](https://api.archbee.com/api/optimize/1Jp0iolWv35Z_mkI7oUiT/dzbqB_wAhmvn4y4ZE86hg_screenshot-2026-08-06-at-21935-pm.png)





![](https://api.archbee.com/api/optimize/1Jp0iolWv35Z_mkI7oUiT/t-6Z_7FLjvPTui6F9zUXu_screenshot-2026-08-11-at-51837-pm.png)

:::hint{type="success"}
![](https://api.archbee.com/api/optimize/1Jp0iolWv35Z_mkI7oUiT/kO1-UQY-VslLIUSjWLigM_image.png)
:::



### Summary&#x20;

By consolidating known and pseudonymous customer data from diverse sources—such as CRM systems, behavioural events, and device IDs—RTCDP creates a single, actionable view of your customers.   AEP Provides a central repository for all of your audiences regardless of how they're created.  Enrich and segment unified profiles in real time with the segmentation service and leverage AI-powered insights to fuel your workflows, ensuring your campaigns are on target and relevant.



#
::::
:::::



