You can configure the AI Assistant feature to be available to specific groups during content creation. You can bring your own AI Provider API key. Provider keys are encrypted, never exposed to users, and validated before activation. This feature automatically includes server-side access controls, usage limits, content guardrails, and an indication whenever AI-generated content is inserted.
Sections in this article:
Considerations
- Supported AI Providers: Currently, only specific non-image-generating models from OpenAI and Gemini are supported.
- One Model per Digital Workplace: You can only configure one model to use throughout the digital workplace.
-
Usage limits: The following limitations are in place to safeguard usage.
- Max requests per user per hour: 20
- Rate limit window: 1 hour
- Max question length: 2,000 characters
- Max context length: 50,000 characters
- Max output tokens: 1,000 tokens
- Supported Content Types: Supported content types include blog articles, calendar events, forum topics, and wiki articles.
Who can do this?
- Enterprise administrators
- Workplace administrators
Configure the AI Assistant feature in content creation
- On the Userbar, select
Control Panel.
- Under Optimization, select AI Assistant.
- Configure the following fields:
- Provider: Select an AI provider. Currently, only OpenAI and Gemini are supported.
-
API key: Paste the API key from your AI provider into the field and select Test key to validate it. OpenAI keys must start with
sk-. Gemini keys must start withAIza. If you cannot find your API key, please contact your IT administrator. - Model: Select an AI model. Only text-based models are available.
- Groups allowed to use AI Assist: Select which groups should be able to use the AI assistant.
- Acknowledgement: Select the checkbox to agree to the acknowledgment statement.
- Select Save Changes.