AI Policy

Approval-first design

Kahvor may draft replies, summarize threads, identify decisions, suggest follow-ups, and organize work. Important actions should remain reviewable and approved by a user or authorized workspace member.

Context visibility

Kahvor aims to keep source context visible so users can understand why a recommendation was made and where the underlying information came from.

User responsibility

AI output can be incorrect, incomplete, or outdated. Users are responsible for reviewing content before sending, relying on, delegating, scheduling, or archiving work.

Restricted uses

Kahvor may not be used to generate unlawful content, impersonate people deceptively, make automated high-impact decisions without appropriate review, or process data without authorization.

Model and provider processing

Kahvor may use third-party AI infrastructure to provide AI features. We design integrations to transmit the context needed for the requested feature and to protect customer data under applicable agreements.