Reference

AI agents and the debugging toolkit

How Adobe's agent layer sits on top of AEP, which tool to debug each product with, and the partner systems that most implementations connect to.

AI Agent Orchestration

AI Agent Orchestration - Adobe Experience Platform Agent Orchestrator.
sits on top of AEP

Reasoning engine, a knowledge base, and a conversational interface (AI Assistant).

this space is evolving quickly, confirm the current agent lineup and licensing with Adobe/your CSM before scoping a project around it

Agents on your board

Data Insights Agent (surfaces analysis inside CJA)
Data Engineering Agent (schema/pipeline assistance across RTCDP, AJO, and CJA)
Workflow Optimization Agent
Product Support Agent (project/process assistance)
Brand Concierge, a conversational, agent-driven customer-facing experience layer

Current as of Oct 2026. Adobe Experience Cloud is now branded Adobe CX Enterprise, and CX Enterprise Coworker (GA June 2026) runs multi-step work from a goal. Agent Orchestrator powers the agents inside existing apps, and AI Assistant is the chat interface to them. Adobe currently lists these agents:

Audience Agent: Builds and tunes audiences from plain-language prompts.
Data Insights Agent: Answers data questions and builds visuals in Analysis Workspace (CJA).
Journey Agent: Analyses and optimises multi-step journeys in AJO.
Experimentation Agent: Sets up and reads experiments to find growth opportunities.
Content Advisor Agent: Finds content and creates on-brand variants.
Brand Experience Agent: Site modernisation, faster forms and help with Cloud Manager build failures (AEM).
Brand Governance Agent: Checks brand policies, permissions and rights.
Product Support Agent: Troubleshoots issues and raises support tickets.
Site Optimization and LLM Optimization Agents: Find and fix website issues, and track how the brand appears in AI-generated answers.
Product Advisor Agent: Product discovery for shoppers, used with Adobe Brand Concierge.
Adobe Marketing Agent for Microsoft 365 Copilot: Brings AEP marketing insights into Microsoft 365 apps.

Source: Agentic AI in CX Enterprise applications. See What changed for all renames.

Orchestration pattern to design toward

Conversational entry point (AI Assistant) captures user/marketer intent in natural language
Reasoning engine decomposes intent into steps and selects which agent(s) — Adobe or third-party — should execute each step.
Agents act against real AEP data/schemas/segments with human-in-the-loop approval for high-impact actions (e.g., large audience activation, spend changes)
Outcomes are logged back into AEP/CJA for measurement and further refinement.

Debugging toolkit

Product	Where to look when something breaks
AEP	Data Prep validation preview, Dataset monitoring UI, Flow Run history, Query Service (SQL against the data lake)
RTCDP	Segment membership preview, Audience portal export logs, Destination “Diagnostics” tab
Launch	Debug mode, Adobe Experience Platform Debugger (Chrome extension), Assurance (real-time event inspection & validation rules)
Adobe Analytics	Debugger extension, Report Builder for spot-checks, Data Feeds for hit-level audits
AJO	Journey simulation/preview, Event/Profile inspection in AEP, delivery/bounce logs per channel
CJA	Data View “Component” validation, Connection ingestion status, cross-check against source Data View
Target	at.js/alloy.js console logging, Preview links (QA mode), Activity Diagnostics panel
AEM	AEM Developer Console, error logs (author/publish), Dispatcher cache-header inspection
Marketo	Lead activity log (Smart Campaign trace), Sample Test Lead, Munchkin debugger, API call logs

Cross-product integrations

CRM (Salesforce, Microsoft Dynamics, HubSpot)
ESP add-ons and ad platforms (Meta, Google Ads, The Trade Desk)
Data warehouses/BI (Snowflake, BigQuery, Databricks, Power BI, Tableau)
Consent Management Platforms (OneTrust, TrustArc, Adobe’s own CMP)
Commerce platforms (Adobe Commerce/Magento, Shopify, SAP Commerce)
Customer support (Zendesk, ServiceNow)
Identity resolution partners (LiveRamp)