Privacy & Responsible Governance

Governance,
Not Surveillance.

Tracelet has visibility into how employees use AI. That responsibility shapes how the product is built. We believe that responsible AI governance and employee surveillance are not the same thing — and we’ve engineered the platform to keep them apart.

Tracelet gives organisations governance without turning AI adoption into surveillance.

Why This Matters

Useful. And trustworthy.

To be useful, Tracelet must observe AI activity. To be trustworthy, it must observe responsibly.

The product's privacy posture is not an afterthought — it is part of the core design. Organisations that buy Tracelet are buying a tool that their employees can trust.

Design Principles

Privacy-Aware Design Principles

The platform is designed around seven privacy principles. These are defaults, not optional features.

01

Classify content instead of exposing full prompts where possible.

Most policy decisions only need to know the category of content, not its full text.

02

Capture metadata when full content is not required.

User, timestamp, destination, and policy outcome are sufficient for most evidence purposes. Full content is opt-in only.

03

Separate business and personal AI usage.

The browser extension distinguishes personal from business sessions. Personal usage is treated differently.

04

Exclude personal usage from productivity analytics.

Productivity metrics only consider business AI activity.

05

Configurable retention policies.

Customers control how long data is retained. Custom retention is available on the Enterprise package.

06

Role-based access to sensitive evidence.

Sensitive content access requires elevated permissions and is itself audited.

07

Redaction of sensitive content in reports.

Where sensitive content must be referenced, it is redacted by default.

Data Handling

What We Capture vs. What We Don’t

By default, Tracelet operates on metadata and classifications — not content. The platform determines whether a policy was triggered and what category of information was involved, without recording the full text of what an employee wrote or received.

Personal AI usage is separated from business AI usage and kept out of governance reporting entirely.

Default — always on
Content classification
Policy outcome (allow / warn / block)
User & timestamp metadata
Destination service
Event type
Opt-in only — explicit config
Full prompt content
Full AI response content
Verbatim tool call arguments

Full-content logging is a deliberate choice the organisation makes with documented data handling controls. Never on by default.

Access Model

Role-Based Visibility

Tracelet’s access model reflects how organisations actually work.

Security administrators

See policy posture, violations, and the ability to drill into specific events for incident investigation.

Compliance officers

See framework-mapped evidence, exception registers, and audit-ready exports.

Department managers

See team-level patterns relevant to their scope — not individual content.

Executives

See organisation-wide adoption, risk, and policy posture summaries.

Individual employees

Can see their own activity. They are not opaque to themselves.

Sensitive content access

Requires elevated permissions, generates an additional audit record, and is reviewable by the customer's compliance team.

Retention

Retention Controls

Retention is configurable. Customers can set different retention windows for each data type. Custom retention policies and extended retention are available on the Enterprise package.

Policy evaluation metadata
Configurable
Sensitive data event evidence
Configurable
Full-prompt logs (if opted in)
Configurable
Audit logs of platform actions
Configurable
For Employees

If Your Employer Deploys Tracelet

Here’s what that means in practice.

Personal browser AI use

Not captured for productivity analytics. The extension distinguishes personal from business sessions.

Your own activity

Visible to you, not just to administrators.

Sensitive content

Not accessible to your manager by default. Access requires elevated permissions and generates an audit trail.

Productivity metrics

Describe how AI changes work — not how individuals are ranked.

Tracelet is built to enable AI adoption, not to police it.

Privacy Package

Want the Full Privacy
& Data-Handling Whitepaper?

Request our full privacy and data-handling whitepaper — including our data processing model, retention architecture, and sub-processor list.