When building an AI-infused product, it’s essential to pay close attention to how you explain the features to your users. According to best practices promoted by Google, you should:
- Explain the benefit, not the technology: Focus on the user needs you are satisfying, rather than how the system works under the hood.
- Set the right expectations: All AI systems have blind spots and will eventually fail. Think carefully about the types of errors they might make and create policies to help users gain control in those situations.
- Explain for understanding, not completeness: Keep in mind that explanations should help users make decisions and be able to move forward.
- Go beyond in-the-moment explanations: For further explanations, provide resources outside of product flows.
I find these recommendations invaluable – something Product Managers, Designers, and Data professionals should always keep in mind.
We tend to fall in love with shiny new technologies, but our focus should always be on the value we provide to users.
Apple is especially good at applying these sorts of practices. In fact, they take it to an extreme by adding their own style to their communications – avoiding the term “AI” and showcasing all of the use cases that are improved by ML (as they insist on calling it).