We will continue to evolve this pattern and the usage guidelines based on insights from product teams and other experts across SAP. Please submit feedback for improvements through the Design System Jira (select “Business AI” in the Stakeholder field), and check our AI acknowledgment Jira ticket and AI patterns roadmap for the latest updates.
Current version 1 has been reviewed by the SAP design system team and the following activities are in progress: accessibility review of specifications and SAPUI5 and SAP Fiori technical planning.
Intro
AI acknowledgment brings essential information and announcements about a present AI service to the users’ attention in a standardized way. Its primary aim is to educate users about various aspects of AI-enabled applications, including information about potential risks and opportunities and the service’s limitations. Doing so establishes a baseline for user expectations and lays the foundation for building trust.
The AI acknowledgment is an integral part of the broader landscape of “messaging patterns” that emerge within AI services.
Example: AI Acknowledgment Popover on new SAP CX AI Toolkit capabilities
Usage
High-Stakes Situations
AI acknowledgment is highly recommended in situations, where educated user decisions on AI created results are critical. In other words, the risks associated with such an action are usually high.
• Medical Advice / Diagnosis
• Legal Advice
• Educational content
• News Platforms and Fact-checking situations
Low-Stakes Situations
AI acknowledgment may be optionally used in situations where the risks associated with applied actions are usually low and can be easily reverted at any time. Other patterns such as AI explainability and/or AI notice may already provide enough visual cues to inform the user of possible AI involvement in the provided content.
• Entertainment
• Recipe / Lifestyle suggestions
• Generic communication responses and greetings
Do
- Inform the user about a present AI service or function.
- Educate the user on important properties of the AI system.
- Notify the user about recent changes to the AI service.
- Message the user about upcoming or current maintenance.
Don't
- Don’t message the user about failed or successful AI processes using the AI acknowledgment pattern.
- Don’t send marketing or advertisement information about the AI service.
- Don’t use the AI acknowledgment pattern for any other system-related information.
Anatomy
Quick prompts pattern extends the following components to enable new AI-specific interactions:
• Dialog / Message Box
Example: Layout Structure - AI Acknowledgment
Structure
- Container*: Dialog / Message Box
- Image / Illustrated Messages (optional): Image representing the AI service Disclaimer: There are currently no specific AI-Illustrations available. Please request new images if you need specific illustrations for you scenario.
- Intro / Primary Message*: What is the message?
- Headlines (Optional): Descriptive headline for provided paragraph.
- Reason / Key Aspects*: Why do I get this message?
- Details / Link*: Where can I find further details?
- Check-Box Option: “Don’t show this message again”
For more information see utilised/reused components in this pattern: Dialog, Message Box, and Illustrated Messages.
Types
Recommended message texts
Footer-Actions
The message type and offered options in the dialog have an impact on which actions are recommended to be displayed in the footer.
Text Only Message:
1. Tertiary: “Close” button
Message with “Don’t show again” option:
2. Primary: “OK” button
Message with explicit user acknowledgment:
3. Modal: options “OK” and “Cancel”
4. Semantic Modal: options “Accept” and “Decline”
Example: Footer Action Variants
Behavior and Interaction
Trigger and Events
The AI acknowledgment – dialog might be launched once the user enters a screen that provides functionalities or carries results of AI based on the current need of communication. This might include events such as:
New AI Feature:
Onboarding / welcome message
Acknowledgments and/or Disclaimer
Service Updates:
Service update information
Service unavailable:
System maintenance message
Error or temporary out of service
“Don’t Show Again” Option
You can provide a “Don’t show this message again” option in the acknowledgment popover to prevent the recurrence of identical messages. If this option is activated, the message is no longer displayed in future sessions in the same application locally, or for the same AI service globally. It does not prevent other messages from showing up. We recommend providing a dedicated location where users can retrieve this information again at any time.
Proposed locations:
SAP Help Portal
My Home Settings (User Profile)
Example: Don’t show this message again
When to Show Again
The popover can be re-surfaced if the status or properties of the AI service have changed significantly – even if the user has selected “Don’t show this message again”. We recommend displaying the AI acknowledgment again in the following scenarios:
- On a first-time encounter of an additional, new AI service in a given application.
- A newly added AI capability, feature, or service.
- A decrease in the overall reliability/output quality of the AI system.
- When the AI service is unavailable because of system failure, maintenance, or high load.
- A major change or update to the AI service and/or design of the function.
This behavior might be configured and personalized by the customer. Customers might also choose to communicate the related information through different means.
Responsible AI
Make consent explicit
If the user consents to privacy statements, terms of use, and other policies required before accessing the AI service or using its features, utilize checkboxes as outlined in the AI acknowledgment pattern to capture explicit consent. Allow users to easily revoke their consent at any time.
Provide easy access to information
Clearly indicate links within the AI acknowledgment so that users can access and review the full text of privacy statements, terms of use, and policies before consent.
Communicate clearly
Use clear, simple language in privacy statements, terms of use, and policies to ensure users understand the terms and conditions they agree to. Avoid technical jargon or legal language that may confuse users.