What AI features do
- Transcribe or summarize voice notes and message threads where enabled.
- Retrieve relevant message sources for search questions and operational follow-ups.
- Draft daily briefs, answer cards, owner/action suggestions, and review flags.
- Highlight uncertainty, low-confidence transcripts, or source gaps where product features support that behavior.
Human review
Retena is an assistive product. AI output should be reviewed against source messages before important operational, financial, legal, safety, or customer-facing decisions.
Data sent to AI providers
When AI features are enabled, Retena may send prompts, message excerpts, transcripts, metadata, source references, and outputs to AI or transcription providers needed to generate the requested result.
We aim to limit inputs to the context needed for the feature and to avoid sending unnecessary secrets, full identifiers, or unrelated private content.
Training and improvement
Retena does not sell customer content or use it for advertising. We do not intentionally use customer message content to train third-party foundation models unless the customer separately opts in or a signed agreement permits it.
We may use aggregated, de-identified, redacted, or synthetic signals to evaluate quality, improve retrieval, reduce errors, and improve product experience.
Restricted AI uses
Do not use Retena as the only decision system for emergency response, medical advice, legal advice, employment, housing, credit, insurance, financial eligibility, public benefits, or other regulated high-impact decisions.
Voice notes and transcripts
Voice transcription can mishear names, amounts, dates, locations, or domain-specific vocabulary. Source audio or original messages should be reviewed before acting on sensitive details.