What DemoTime feedback looks for—and what it cannot prove.

Feedback dimensions

• Clarity: whether the main point and requested action are easy to identify.
• Structure: whether context, reasoning, evidence, and next steps appear in a useful order.
• Specificity: whether claims are supported by concrete actions, constraints, owners, or outcomes.
• Tone: whether wording fits the relationship, purpose, and level of pressure.
• Delivery: pace, pauses, rhythm, pronunciation, and other signals that can affect intelligibility.
• Conversation handling: whether the response answers the question and creates a workable next turn.

How to interpret a score

A score is a practice signal for the response and scenario being reviewed, not an objective measure of a person. Different prompts, microphones, network conditions, accents, or AI behavior can change the result. Compare attempts within the same scenario and use the written evidence behind a score.

Limitations and human judgment

Automated feedback may miss context, misunderstand speech, overvalue one communication style, or suggest wording that is unsuitable for a culture or workplace. Keep useful suggestions, reject unsuitable ones, and ask a trusted human reviewer when consequences are significant.

Voice and training data

DemoTime processes raw audio as needed for the live conversation and transcription and does not retain raw audio after real-time processing is complete. Voice, transcripts, practice content, and feedback are not used to train general-purpose AI models unless a user separately and expressly opts in.