Work with AI ai-features

• View Journeys - View insights into journeys directly in AI Assistant
• Manage Journeys - Create new journeys directly in AI Assistant
• View Segments - View insights into audiences and search existing audiences
• Manage Segments - Create new audiences directly in AI Assistant
• View Journey Events, Data Sources and Actions - Required for the Create skill to search journey events and custom actions

• Journey Fallout Analysis - Identify where and why customers drop off during journeys, detect disengagement patterns
• Audience Overlap Detection - Analyze audience overlap across multiple journeys to prevent fatigue from over-targeting
• Schedule Conflict Detection - Identify timing conflicts between scheduled journeys targeting the same audience
• Operational Insights - Get prompt-based insights like “show me all live journeys” or “which audiences are used in more than X journeys”

• “Perform a fallout analysis for journey [Journey Name]”
• “Are there any scheduling conflicts for journey [Journey Name]?”
• “Show me audience overlap conflicts for journey [Journey Name]”
• “Which audiences are used in more than 5 journeys?”

• Natural Language Journey Creation - Describe your desired journey and have it created automatically
• Event- and Audience-Based Starts - Create event-triggered, audience-based, business-event, or audience qualification journeys
• Conditional Logic - Build split paths based on customer attributes or behavior
• Multi-Channel Messaging - Add email, push, and SMS actions
• Scheduling - Configure start dates and timing between steps

• “Create a journey that starts when a customer makes a purchase online and sends a thank you push notification.”
• “Build a journey targeting my day hikers audience with three emails over two weeks, starting 12/20.”
• “Create a journey that starts when a user enters my store location and follows up based on whether they have a valid email address.”

• Performance Analysis - Clear view of what happened in experiments
• Insights Generation - Explanation of why results occurred
• Opportunities Discovery - Guidance on next actions to take
• Content Analysis - Examine messaging elements to understand why certain treatments outperformed others
• Recommendation Generation - Suggest new treatments or adjustments based on insights

• “What experiments are running for [Campaign Name]?”
• “For my [Experiment Name], what treatment is leading?”
• “What did we learn from [Experiment Name]?”
• “What do you recommend I do next after this experiment?”
• “What common patterns are emerging from recent tests?”

• View Experiments - View insights into experiments in AI Assistant
• Manage Experiment Metadata - Create new experiments in AI Assistant

Note: Available with Journey Optimizer Experimentation Accelerator license.

Adobe Journey Optimizer can automatically analyze Decisioning rules and ranking formulas expressed in PQL syntax, and suggest simplifications that preserve the original logic. When a simplification is found, a red Optimize indicator appears next to the rule or formula, opening a side-by-side comparison of the original and AI-suggested expressions, with a downloadable analysis to validate that both behave identically.

Key Capabilities

• Logic-preserving simplifications - The AI suggests a shorter expression that returns the same result on simulated profiles.
• Validation report - Download an analysis (TSV) showing how each simulated profile is evaluated against both versions before applying the change.
• One-click apply - Replace the original PQL with the optimized version directly from the Optimize window.

Eligibility

Only rules and ranking formulas whose PQL expression is larger than 2 KB (UTF-8 encoded) are targeted for analysis, smaller expressions are not analyzed.

Permissions

This capability uses the same generative AI access controls as Generate Content. Users must be granted the Generate Content permission on the AI Assistant resource. Learn more about Generate Content access

Optimize Decisioning rules | Optimize ranking formulas

Experiment Accelerator helps you run experiments faster with AI-driven insights and recommendations, identifying winning content variations more quickly.

Key capabilities:

• Generate multiple content variations automatically
• Receive AI recommendations for experiment design
• Get early indicators of performance trends
• Accelerate time to statistical significance

Learn more about Experiment Accelerator

Use AI to predict the optimal time to send each message based on individual customer behavior patterns, maximizing engagement.

How It Works

Send-Time Optimization analyzes historical engagement data (opens and clicks) to predict when each customer is most likely to engage with your messages. The system automatically schedules delivery within your specified time window.

When to Use It

table 0-row-2 1-row-2 2-row-2 3-row-2 4-row-2
Best For	Not Recommended For
Marketing campaigns and newsletters	Time-sensitive operational messages (order confirmations, password resets)
Promotional messages	Urgent notifications (flight delays, emergency alerts)
Educational content	Event-based messages with specific timing requirements
Engagement campaigns

Learn more about Send-Time Optimization

Transform static image designs (JPEG, PNG) into editable HTML email templates using AI-powered conversion technology.

Learn more about Image to HTML

Integrate with Adobe GenStudio for Performance Marketing to create AI-powered email content and import templates into Journey Optimizer for orchestration. Export Journey Optimizer templates to GenStudio, generate variations with AI, and bring them back for deployment. (Limited availability, email channel only.)

Learn more about GenStudio

Evaluate how well your content aligns with your brand guidelines using AI-powered scoring that measures tone, voice, and messaging consistency.

Learn more about Brand Alignment

AI models for decisioning ai-decisioning

Create intelligent ranking models that rank offers by conversion rate (conversions ÷ impressions), automatically showing each customer the offer most likely to convert.

Model Types

• Auto-optimization - Learns from the overall, non-personalized performance of your offers to automatically improve conversion over time. A good fit when offers change often, since the model retrains roughly every 6 hours.
• Personalized optimization - Uses customer profile attributes, behavior, and audience membership to predict the best offer for each individual. Choose this when you need a different ranking per customer rather than one overall winner.

Requirements

Minimum data requirements differ by model type:

• Auto-optimization - At least 2 offers with 100+ display events and 5+ click events each within the last 14 days. Offers below this threshold are treated as new and only served through exploration traffic.
• Personalized optimization - Uses a rolling 30-day window. Adobe recommends at least 1,000 impressions and 100 conversion events per offer per week; by default, offers with fewer than 1,000 impressions or 50 conversions won’t get a model trained for them. Up to 5 audiences can be selected to train a single model.

Learn more about AI models for decisioning | Create AI ranking models

AI Knowledge Reference

This section contains structured knowledge intended to support interpretation, retrieval, and question answering related to this topic.

For complete understanding, this information should be combined with the documentation on this page. Neither source is intended to stand alone; the page describes the feature, while this section provides additional context that helps disambiguate terminology, intent, applicability, and constraints.

• TL;DR: This page is an overview of the AI and machine learning features across Journey Optimizer, including CX Coworker, AI Assistant, AI agents, AI-powered content generation, integrations and tools, Send-Time Optimization, AI decisioning, and content experimentation.

Intents:

• Understand what CX Coworker is, how it relates to AI Assistant, and which skills it provides for journeys, loyalty, and content management
• Understand what AI Assistant can do and how to access it
• Learn which AI agents are available and what permissions each requires
• Generate content and message variations with AI across supported channels
• Predict the best send time per customer with Send-Time Optimization
• Rank and personalize offers with AI decisioning models
• Optimize Decisioning rules and ranking formulas with AI-suggested simplifications

Glossary:

• CX Coworker: Adobe’s conversational experience that integrates with business applications, automating customer experience and marketing workflows across Adobe applications; you describe a business goal in natural language and Coworker plans, executes, and validates the work (product-specific)
• Journey Create (CX Coworker skill): A CX Coworker skill for journeys that builds journeys with natural language prompts (product-specific)
• Channel Content Create (CX Coworker skill): A CX Coworker skill for journeys that generates channel-specific message copy and designs (product-specific)
• Journey Analyze (CX Coworker skill): A CX Coworker skill for journeys that diagnoses journey drop-offs, scheduling overlaps, and execution anomalies (product-specific)
• Loyalty Challenge Management (CX Coworker skill): A CX Coworker skill for loyalty that creates and configures loyalty challenges (product-specific)
• Loyalty Insights (CX Coworker skill): A CX Coworker skill for loyalty that queries loyalty points, tier distributions, and revenue metrics (product-specific)
• AI Assistant: The conversational guide to Journey Optimizer that answers product-knowledge questions and, in Beta, surfaces operational insights about journeys (product-specific)
• Journey Agent: An AI agent offering two skills in AI Assistant, Analyze and Create, for optimizing or building journeys from natural language (product-specific)
• Experimentation Agent: An AI agent that analyzes and manages digital experiments across websites, emails, push messages, and applications (product-specific)
• Generate Content: AI-powered content generation for subject lines, body text, images, and message variations (product-specific)
• Auto-optimization: A ranking model type that learns from overall, non-personalized offer performance and retrains roughly every 6 hours (product-specific)
• Personalized optimization: A ranking model type that uses customer profile attributes, behavior, and audience membership to predict the best offer per individual (product-specific)
• Send-Time Optimization: An AI capability that predicts the optimal send time per customer from historical engagement data (product-specific)
• Journey Optimizer MCP server: A capability that connects Journey Optimizer to MCP-compatible AI applications for plain-language queries; all operations are currently read-only (product-specific)

Guardrails:

• Content Management (CX Coworker) is powered by 15 read/write MCP tools.
• AI Assistant requires agreement to the Adobe Experience Cloud Generative AI User Guidelines before use.
• Operational Insights are in Beta, are currently only available for Journeys, and reflect data from your current sandbox.
• Send-Time Optimization is only available for Email and Push actions in journeys; your organization needs at least 30 days of history using those actions before enabling it.
• AI Content Generation is only available for the Email, Push, Web, and SMS channels.
• Brand asset uploads support PDFs, images, or ZIP files (max 50 MB) (hard limit).
• Custom templates support up to 8-10 images.
• Auto-optimization requires at least 2 offers with 100+ display events and 5+ click events each within the last 14 days; offers below this threshold are treated as new and only served through exploration traffic.
• Personalized optimization uses a rolling 30-day window; Adobe recommends at least 1,000 impressions and 100 conversion events per offer per week (recommended); by default, offers with fewer than 1,000 impressions or 50 conversions do not get a model trained for them.
• Up to 5 audiences can be selected to train a single personalized optimization model.
• AI-powered rule and formula optimization analyzes only rules and ranking formulas whose PQL expression is larger than 2 KB (UTF-8 encoded); smaller expressions are not analyzed.
• AI-powered rule and formula optimization requires the Generate Content permission on the AI Assistant resource.
• The Journey Optimizer MCP server currently supports read-only operations only.
• GenStudio for performance marketing integration is in Limited availability and email channel only.
• The Experimentation Agent is available with the Journey Optimizer Experimentation Accelerator license and requires View Experiments and Manage Experiment Metadata permissions.

Terminology:

• Canonical name: Send-Time Optimization — Acronym: n/a — variants: send-time optimization
• Synonyms: “Generate Content” = “AI content generation”
• Do not confuse: “Auto-optimization” (learns from overall, non-personalized performance) ≠ “Personalized optimization” (predicts the best offer per individual)
• Do not confuse: “Journey Analyze skill” (optimize existing journeys) ≠ “Journey Create skill” (build new journeys from prompts)
• Do not confuse: “AI Assistant” (conversational guide) ≠ “AI Agents” (specialized agents for deep analysis and recommendations)
• Do not confuse: “Journey Analyze skill” and “Journey Create skill” (AI Assistant’s Journey Agent skills) ≠ “Journey Analyze (CX Coworker skill)” and “Journey Create (CX Coworker skill)” — the page names these separately: the Journey Agent skills operate within AI Assistant, while Journey Analyze and Journey Create are CX Coworker skills for the Journeys capability.
• Do not confuse: “CX Coworker” (Adobe’s evolution of AI Assistant, described in the CX Coworker section) ≠ “AI Assistant” (the conversational guide it is succeeding)

FAQ:

• Q: How does CX Coworker relate to AI Assistant? — CX Coworker is Adobe’s evolution of the AI Assistant conversational experience; organizations are being transitioned from AI Assistant to CX Coworker on a rolling basis, and some organizations may retain access to AI Assistant until they are enabled for CX Coworker.
• Q: What permissions do I need for AI features? — Generate Content requires the Generate Content permission; AI Assistant product knowledge requires agreement to the Adobe Generative AI User Guidelines; each agent requires its listed View/Manage permissions.
• Q: Is AI-generated content always accurate? — No; always review AI-generated content for accuracy and brand appropriateness, and use the thumbs up/down feedback tools.
• Q: Which channels support AI Content Generation? — Only the Email, Push, Web, and SMS channels.
• Q: What are the requirements for AI ranking models? — They differ by model type; Auto-optimization and Personalized optimization have distinct minimum interaction-data thresholds described in Requirements.
• Q: How do I get access to these features? — Most AI features are included with Journey Optimizer; some capabilities such as Send-Time Optimization or AI Agents may require enablement by Adobe, so contact your Adobe representative.