---
title: Lab 1 - CX Coworker
slug: sydney-cxo/lab-1-cx-coworker
docTags: 
createdAt: 2026-06-12T03:29:50.965Z
---

In this hands-on lab, you will step into the shoes of a Luma marketing team navigating tax refund season. Using Adobe Experience Platform CX Coworker - you will learn how to ground AI in your organisation's own business context, then use it to analyse customer behavior, build policy-aware audiences, and launch personalised journeys across Customer Journey Analytics, Real-Time CDP and Adobe Journey Optimizer.

## Goals

After completing this lab, you will walk away with the following knowledge:

- How to upload and activate business context in CX Coworker so every AI agent speaks your organisation's language
- How to use the Co-worker to surface actionable trends from Customer Journey Analytics without manual analysis
- How to build insight-driven, policy-aware audiences in Real-Time CDP using natural language prompts
- How to generate and validate personalised journeys in Adobe Journey Optimizer
- How to monitor AI Assistant usage across your organisation to support governance and adoption

## Prerequisites

No prior experience with agentic AI or CX Coworker is required. To get the most out of this lab, it helps to have:

- Basic familiarity with Adobe Experience Platform (AEP) navigation and sandboxes
- A general understanding of what Real-Time CDP, Adobe Journey Optimizer, and Customer Journey Analytics are used for — deep expertise is not required
- Access to the lab sandbox environment — credentials are provided on your lab card
- A Chrome browser with a stable internet connection - wifi details provided on check-in

## The Luma Scenario

Before you start any exercise, read through Luma's situation. Every prompt you write and every decision you make in this lab should be grounded in this context.

### Who is Luma?

Luma is a premium outdoor and athletic apparel brand with a large e-commerce presence. Their customers range from casual weekend shoppers to serious athletes. The marketing team runs campaigns across email, in-app, and web channels, all orchestrated through Adobe Experience Platform.

### The Business Situation

It is early April. Luma's marketing team knows tax refund season is a golden opportunity - customers receive an unexpected windfall and many are in a spending mindset.

But not all customers are the same. Luma has identified three distinct mindsets among their July shoppers:

- **Treat-Yourself Shopper** - browsing premium products, higher average order value. Discounts feel cheap to them.
- **Budget-Conscious Shopper** - high engagement with promotions, lower AOV. Frequency of messaging matters.
- **Value Seeker** - interested in bundles, durability, long-term investment. They want to feel smart about their purchase.

**The challenge: **reach each mindset with a relevant message, without over-messaging, and without discounting to customers who would buy at full price anyway.

### What You Will Build

1. Upload Luma's business context so every agent speaks Luma's language
2. Analyze April website behavior to confirm the tax-season trend
3. Build a "High Disposable Intent" audience - these are customers primed to buy
4. Create a journey that routes each shopper to the right message and suppresses recent purchasers
5. (Stretch) Monitor your team's AI usage and explore custom agents

### Technical Architecture

:::hint{type="danger"}
NEEDS TO BE UPDATED FOR CO-WORKER 
:::

Here is a simple architecture diagram that can help you understand the technical stack behind the experience you are about to try out. Adobe Experience Platform Agents use intelligence, speed, and scale to execute complex decision-making and problem-solving tasks. All of this is made possible by the five core components below

1. **Conversational interface:** AI Assistant is an immersive user experience through which Experience Platform Agents and other AI capabilities can be accessed.
2. **Reasoning engine:** The intelligence core that translates natural language into a series of agentic executions. The reasoning engine is designed to interpret the goals and constraints from a user’s inputs and formulate a dynamic plan to achieve that goal.
3. **Agents:** The building blocks required to execute predefined, task-based agentic skills.
4. **Knowledge base:** A central repository of data sources and services that provides grounding for the reasoning engine and supports role-oriented prompts.
5. **Agent composer:** A design-time interface that enables configuration and customisation of pre-built Experience Platform Agents.



![](https://main--adobe-labs--adobe-showcase.aem.page/events/summit26-sydney/cxo/l815/media_1f0729a9b78d66b34dff6b98a2648dd81155e1690.png?width=750\&format=png\&optimize=medium)