Definition

Summaries: A natural language summarization of a data set.

Insights: A natural language explanation of patterns and relationships in a dataset.

Recommendations: Prescriptive guidance users can take to make a measurable impact on their data.

Requirements

1. Attribution pattern [link to UI]

Deliver design to communicate content is authored by AI, in words or by symbols.

2. Contextual data pattern [link to UI]

Deliver design to give the user a way to navigate to the data that informed the summary, insight, or recommendation.

3. Explainability patterns [link to template]

If the contextual data pattern is not present, then deliver a design to explain to users the inputs behind and output and link to the Qualtrics data privacy policy.

4. Ethical disclosure pattern [link to UI]

Deliver design to warn users of potential inaccuracies in an AI output.

Case studies