Design Workbook
The Psychology of AI User Experiences
Back to start
6/8
6. How human should it be?
Should your product have a humanized name?
Refer back to your mental model from Section 2. A companion may warrant a name; a tool probably doesn't.
What is the name of your product/feature?
What tone should the AI convey?
Tone isn't just conveyed through chat. It comes through in microcopy, notifications, error messages, animations, voice, etc. Is the tone neutral, mechanical, professional, friendly, empathetic, a blend or something else?.
How will you frame its level of competence?
This shapes user expectations. A beginner framing invites patience; an expert framing demands accuracy.
Can users control how human the AI feels?
Can they adjust the name, tone, personality, avatar etc.? If so, how?
Will you include other humanizing elements?
Examples include typing indicators, conversational fillers, humor, and memory of past interactions.
Is there a risk of making your AI product/feature feel too human? What is it?
Examples include overattachment, inflated expectations, uncanny valley etc.
NOTES: Explain why you made your choices.
BackNext