Confidence by Spotify launches product intelligence

Confidence is more ambitious than another A/B testing dashboard: it aims to close the loop from product signal to tested action. The strongest pitch is pairing AI assistance with transparent experimentation and human approval.

• –Warehouse-native architecture keeps existing business data and models as the source of truth.
• –Confidence Agent can answer product questions in plain language, generate queries, and expose its sources for review.
• –Feature flags, rollouts, experiments, recordings, and analytics live in one decision loop instead of disconnected tools.
• –Spotify’s experimentation heritage adds credibility, but smaller teams may not have Spotify-scale data maturity to reproduce the same results.
• –Human approval is a meaningful safeguard against blindly shipping AI-generated hypotheses.