Opal AI Platform

Marketing teams juggle content creation, campaign management, experimentation, and data analysis across multiple tools. Every handoff between systems costs time. Writing the same request differently for each tool wastes effort. Maintaining brand consistency when dozens of people produce content across a dozen products is nearly impossible without automation.

Opal is Optimizely’s agent orchestration platform. It sits across Optimizely One and gives you a single, natural-language interface to automate tasks, surface insights, and guide decisions. You describe what you need, and Opal selects the right tools and agents to deliver it.

Opal is an AI layer that runs across every product in Optimizely One. Rather than a standalone application, Opal embeds itself into your existing workflows through Chat, product-level integrations, and purpose-built agents.

Opal provides three core functions:

• Answer questions about platform terminology, configuration, and best practices
• Configure and create content, campaigns, and experiments through natural language
• Automate repetitive tasks so your team focuses on strategy rather than production

AI agents are software systems that act on your behalf. Unlike a simple text generator, agents analyze information, interact with platform data, and adapt based on feedback. Opal classifies its agents into three types:

Agent type	Purpose	Example
Simple assistants	Straightforward tasks like content generation and formatting	Drafting a blog post introduction
Specialized agents	Domain-specific analysis using expert-level context	Reviewing experiment results against your KPIs
Workflow agents	Multi-step processes coordinated across integrated systems	Generating a campaign brief, creating assets, and scheduling publication

Every interaction with Opal follows an eight-step pipeline. Understanding this pipeline helps you write better prompts and troubleshoot unexpected results.

1. User prompt — You enter a natural language request. Opal analyzes your intent, context, and any uploaded files.
2. Instructions — Opal loads the foundational guidelines that define its persona and approach for your organization, including brand voice and compliance rules.
3. Contextual intelligence — Opal integrates platform data, workspace context, and product-specific information relevant to your request.
4. Strategic tool selection — Based on your request, Opal identifies which tools (campaign creation, image generation, web search, etc.) to invoke.
5. Input generation and validation — Opal gathers required parameters such as dates, email addresses, or queries and validates them before execution.
6. LLM interaction — The assembled context, instructions, and parameters are processed through advanced AI models.
7. Execute and refine — Opal takes the planned actions and iterates on results to improve quality.
8. Tailored response — You receive clear, actionable output formatted for your specific context.

Opal is enabled by default for eligible customers, though administrators must grant individual user access through Opti ID.

Opal integrates with the following products:

Product	Opal capabilities
Content Marketing Platform	AI content generation, writing assistance, translation, campaign kits
CMS (SaaS and PaaS)	Inline content suggestions, SEO optimization, image alt text
Web Experimentation	Variation summaries, copy suggestions, experiment review
Feature Experimentation	Flag brainstorming, experiment summarization, test recommendations
Personalization	Variation summaries, copy generation, AI variation development
Optimizely Data Platform	Customer profile summaries, real-time audience suggestions
Analytics	AI-powered insights and anomaly detection
Commerce Connect / Configured Commerce	Product description generation at scale
Content Recommendations	NLP-based content recommendation enhancements
Collaboration	Collaborative AI-assisted workflows
Product Information Management	AI-assisted product data management

Several machine learning features ship with Optimizely products and do not consume credits:

• Stats Engine and Stats Accelerator
• Multi-armed bandit optimization
• NLP-based content recommendations
• Adaptive audiences

Opal uses a credit-based consumption model. Every AI operation — generating a draft, optimizing content, translating text, analyzing images — consumes credits from your organization’s monthly allocation.

Operation	Approximate cost	Notes
Short text generation (headline, meta description)	1 credit	Under 100 words of output
Medium text generation (paragraph, product description)	2-5 credits	100-500 words of output
Long text generation (full article draft)	10-25 credits	500+ words of output
Tone adjustment or rewrite	2-5 credits	Based on input text length
SEO analysis	3 credits	Per page analyzed
Image alt text generation	1 credit	Per image
Translation draft	5-15 credits	Based on word count and language pair
Image generation	Varies	Depends on resolution and count

Credits are allocated at the account level and shared across all users. Administrators can:

• View usage — See total credits consumed, credits remaining, and usage by user
• Set limits — Cap individual user consumption to prevent one person from exhausting the account
• Monitor trends — Track credit usage over time to forecast when additional credits are needed

Credits renew based on your subscription plan. Most plans provide a monthly allotment. Unused credits do not roll over.

Opal requires Opti ID for three reasons:

1. Usage tracking — The credit system attributes consumption to the correct account and user.
2. Access control — Administrators control which users access Opal features and which agents are available to specific roles.
3. Data governance — Content generated by Opal may contain sensitive information. Only authenticated, authorized users can access AI capabilities.

If your organization has not set up Opti ID, Opal features will not appear. See the Opti ID setup guide to get started.

• Instructions shape how Opal responds — from brand voice to compliance rules
• Tools give Opal the ability to take actions in your products
• Agents combine instructions and tools into reusable, purpose-built assistants
• Canvas provides a collaborative workspace for creating and editing digital assets
• RAG connects Opal to your content data for contextually informed responses