Skip to main content

Trust & safety

We analyze behavior, not minds

People practice honestly only when a rough first attempt cannot hurt them. These principles shape how Convrsy works - for the people practising and the organizations that bring us in.

Simulation dynamics versus participant feedback

Counterparts express designed reactions

A simulated counterpart may become guarded, push back or warm up. These are designed simulation dynamics that make the practice realistic.

Participant feedback is grounded in evidence

Convrsy does not claim to detect or score a participant's emotions or internal state. Every observation traces to something that happened in the conversation.

Six principles we build by

  • Practice is private by default

    Individual sessions belong to the person practising. Managers and administrators do not see individual transcripts, recordings or feedback.

  • No performance surveillance

    Convrsy is a practice space, not an assessment tool. People can experiment without penalties or negative scores. Practice data does not feed performance, hiring or disciplinary decisions.

  • Behavior, not emotion

    Feedback describes observable words, choices and interaction patterns. We do not claim to detect anxiety, stress, confidence or any other internal state.

  • Your data stays yours

    Customer content is used only to provide the agreed service. It is not used to train models for other customers, and practice data can be deleted in line with agreed retention settings.

  • Simulations, clearly labeled

    Every counterpart is visibly a simulation. It may reflect user-provided context, but Convrsy does not present it as the real person or claim to reproduce someone exactly.

  • Difficult, not harmful

    Scenarios can include pressure, conflict and pushback — within defined limits. Counterparts may challenge participants, but they do not abuse, humiliate or manipulate them.

Clear boundaries - Convrsy does not

  • Score anyone's anxiety, stress, confidence or empathy
  • Claim that a participant "was anxious" or "became defensive"
  • Give managers access to individual practice transcripts, recordings or feedback
  • Feed practice data into performance, hiring or disciplinary decisions
  • Use confidential customer content to train models for other customers

Feedback language we use instead

  • Asked exploratory questions
  • Acknowledged the counterpart's concerns
  • Interrupted twice
  • Summarized the other party's position
  • Counterpart disclosed additional information
  • Conversation moved from challenge to problem-solving

Data, plainly

Session recordings
Sessions are processed to generate the participant's own review. Individual recordings, transcripts and feedback are not shared with managers, administrators or peers and are protected by strict security controls.
Visibility
Participants see their own conversations, feedback and opportunities to try again. Organizations may see de-identified, aggregated participation and practice insights only for cohorts of 15 or more participants — reporting is withheld for smaller cohorts to reduce the risk of identifying an individual.
Retention and deletion
Practice data is retained according to privacy regulations and agreed settings. Users and organizations can request deletion of their data, subject to applicable legal and contractual requirements.
Client materials
Documents, scenarios and context provided by a client are used only to deliver that client's Convrsy experience. They are not used to train models for other customers.
Model providers
Convrsy uses vetted model providers under agreements that exclude training on customer data. A current subprocessor list is available on request.
Scenario governance
Enterprise scenarios are reviewed and approved by the client before rollout — including their content, counterpart behavior and feedback framing.

Talk directly with the team about deployment, data handling, and scenario governance.

Security or governance questions?