You're exploring a demo tenant. Real data flows, no signup. Editing is disabled.

Responsible AI by Design

Pillars, golden rules and the data sensitivity matrix that governs every approved tool.

Our 4 AI Pillars

Every customer-facing AI output is reviewed by a qualified human before it ships.

We never send PII, financial records, or proprietary code to public models.

If AI helped produce it, we say so — internally and to customers.

Every employee dedicates 1 hour per week to AI literacy. Tracked, not policed.

Never paste customer PII or financial data into a public model (ChatGPT, Claude, Gemini consumer).
Always disclose AI use in client-facing deliverables.
Verify every fact, statistic, and citation a model generates.
Tools handling Tier 3+ data must be on the approved registry.
Report shadow AI use to your team lead — not to HR. We learn from it.
Generative AI cannot make hiring, firing, or compensation decisions.

Tier	Examples	Allowed tools
Tier 0 · Public	Marketing copy, public blog drafts, press releases	Any approved tool
Tier 1 · Internal	Internal docs, meeting notes, project plans	Enterprise ChatGPT, Copilot, Claude Enterprise
Tier 2 · Confidential	Strategy docs, financials, unreleased products	Self-hosted only (Acme GPT, on-prem Llama)
Tier 3 · Restricted	Customer PII, contracts, source code, M&A	Self-hosted + DPO approval required

EU AI Act risk classification is mapped automatically per tool. Tier 3 changes require DPO sign-off.