AI Security & Privacy Overview

AI in Pivot accelerates work without compromising privacy. This page explains how AI features handle customer data and the controls available to admins.

Model Providers & Data Handling

• Minimum necessary context. AI calls send only the prompt/context required to fulfill the request.

• No training on your data. At Pivot, we never train models on customer data. Third-party model providers are contractually restricted from training on, or retaining, your prompts/responses.

• Tenant isolation. Requests are processed in a way that does not mix one customer’s data with another’s. Users can only use AI on data they have access to.

Admin Controls

• Enable/disable AI features org-wide.

• Export and review AI-related events via Audit Logs.

High-Risk / Regulated Data

• Using AI with PHI or similar regulated data requires HIPAA-eligible deployment and a signed BAA.

• Customers are responsible for data classification and for configuring least-privilege access, SSO/MFA, and retention per policy.

Security & Privacy Baseline

Pivot applies the same TOMs described in the DPA to AI flows: encryption in transit and at rest, access controls, logging, and vendor due-diligence.

Ongoing Review

New/updated model providers undergo security, privacy, and data-processing reviews. If a provider changes terms that affect data handling, we will update this page and our sub-processor list