FrontierzDocs

Last updated on

Responsible AI

These four principles explain how Frontierz identifies AI, supports human review, and sets limits on the use of AI Fellows. Use them when deciding how to introduce Frontierz in your organisation.

Every AI interaction is labelled as AI

Sessions, messages, and other platform screens identify the AI Fellow as AI. When introducing it to employees, make clear that they will be interacting with AI, and avoid descriptions that imply it is a human adviser.

Live AI Fellow call screen with the AI Fellow identity visible

Clear identification helps employees understand the experience and judge the responses they receive.

Reviewable evidence for AI outputs

When reviewing an AI response, use the conversation and configuration records to understand the context in which it was given.

For any conversation, authorized administrators can see:

  • The full conversation transcript.
  • The current AI Fellow configuration, including purpose, Knowledge Base, and attached Connectors.
  • Custom Connector executions where relevant, with the data sent and the response received.

Conversation transcript detail showing reviewable evidence

If a user reports an inappropriate response from the AI Fellow, the audit trail supports investigation.

Scope of AI use

AI Fellows support learning, coaching, simulations, and workforce development. They can analyse conversations and produce structured assessments for competency-based programmes.

Purpose editor with limits and scope in the purpose text

AI Fellows are not a substitute for human judgment in high-stakes legal, medical, financial, employment, or compliance decisions. AI Fellows can support training and rehearsal for regulated roles. They are not authorized to make employment, regulatory, or compliance decisions.

When AI Fellow outputs or assessments touch decisions that affect a person's employment, status, or rights, route them through human review.

Data isolation

Company data uploaded to Frontierz is isolated per organization and not visible to other customer organizations.

  • Encryption at rest and in transit.
  • Company-level data isolation by design. Customer knowledge, conversations, and analytics do not mix with another company's data.

Company Profile showing AI Manifesto URL and company context fields

For specific certifications, residency options, retention policies, or deletion guarantees, contact the Frontierz account team. The applicable terms depend on the customer agreement and on configuration that varies by enterprise customer. See Privacy and data for the data-handling details to confirm before commitment.