This guide preserves published Enad setup and compatibility guidance. Enad app is authoritative for current controls. Use generated API/OpenAPI reference and the Event catalog for exact fields, payloads, authentication, and delivery behavior; examples and operational limits below are compatibility guidance until attached to those contracts.

The AI integration makes Enad's AI tools available through either Enad's managed AI service or a supported provider.

What this turns on

An active AI integration enables:

• Draft product descriptions and SEO text with AI Assist.
• Generate individual or bulk translations.
• Create and iterate on campaign images in AI Studio.
• Run bulk AI jobs for text, translations, image alt text, and image processing.
• Reuse prompt templates with Presets.

Most AI operations require an active AI integration. Available tools also depend on the features enabled for the app.

Provider requirements

The provider determines how Enad accesses the AI model:

• Enad uses Enad's managed service and does not require a provider API key.
• Gemini, OpenAI, or Anthropic uses an account and API key from the selected provider.

Configuration

Configuration follows the common integration setup and includes:

• Provider.
• Model.
• Provider API key, except for Enad's managed service.
• Active status.

An app can have multiple AI integrations. Supported AI operations allow selection between active integrations.

Model responsibility

The selected model has a direct effect on output quality. A model that is too limited for the task can produce weak, incorrect, inconsistent, or off-brand results, even when the profile and source data are well configured.

Model capabilities should match the intended work. Translation, long-form product copy, structured SEO text, and image work can require different capabilities.

Production readiness should be assessed through:

• Tests on representative products, languages, and edge cases.
• Comparison against expected quality and tone.
• Review before approval, especially after a model or provider change.
• An AI profile for stable brand, tone, terminology, and content rules.

A profile guides a model but does not compensate for unsuitable model capabilities. Model selection and output review remain the responsibility of the merchant.

How the controls fit together

flowchart LR
    Integration["AI integration<br/>Provider + model"]
    Surface{"Generation surface"}
    Assist["AI Assist<br/>One field"]
    Translation["Translations<br/>One or many languages"]
    Jobs["AI Jobs<br/>Bulk generation"]
    Profile["AI profile<br/>Brand, tone, terminology, rules"]
    Review["Generated result<br/>Review and approve"]

    Integration --> Surface
    Surface --> Assist
    Surface --> Translation
    Surface --> Jobs
    Profile --> Assist
    Profile --> Translation
    Profile --> Jobs
    Assist --> Review
    Translation --> Review
    Jobs --> Review

• The integration chooses the provider and model that produce the output.
• The AI surface decides what to generate and for which records or languages.
• The optional profile adds reusable instructions for how the output should sound and behave.
• Generated results require review and approval.

AI Studio uses Presets for reusable campaign-image prompts. Profiles guide catalog text, translations, and AI jobs; they are not AI Studio presets.

Default AI Studio integration

Default AI studio selects an Enad provider integration automatically for new AI Studio sessions. Other active integrations remain available.

The list of available models can change. The model selector in Enad Admin is the current canonical reference.

Next step

AI profiles keep tone and content rules consistent across supported AI operations.