Executive Decision Pack

SAMPLE-03 · HR policy RAG chatbot (global workforce)
Generated 2026-10-31 14:52 UTC · Owner: Aisha Rahman · Sponsor: Chief People Officer

Scenario

A global manufacturer wants a RAG chatbot over its HR policy library so 40,000 employees across three languages (English, Mandarin, Portuguese) can ask policy questions in natural language. Personal data (salary, medical leave) may appear in retrieved documents. Embedded in Microsoft Teams. Uses an external commercial LLM.

Decision narrative

Initiative SAMPLE-03 (HR policy chatbot) is assessed at ARA overall-residual Low, CALIBRE band Moderate (GII 38.75/100). The retrieval-only design and policy-scoped corpus keep this well within enterprise risk appetite.

Approve. Continue normal monitoring; no additional conditions required beyond the three listed controls.

Top risks driving the decision

R-DA-01 PII exposure via retrieved documents
R-AI-01 Hallucinated policy answers

Conditions for approval

Corpus excludes salary and medical records
Every answer includes source citation
GDPR Article 22 disclosure visible in Teams

Evidence attached

None yet attached.

Remaining evidence gaps

None outstanding.

Actions and owners

ID	Action	Owner role	Priority	Status
SAMPLE-03-A-01	Verify Role-based corpus filtering configuration	HR Tech	Medium	Open
SAMPLE-03-A-02	Verify Citation-required answers behaviour	AI Product Owner	Medium	Open
SAMPLE-03-A-03	Publish GDPR Article 22 disclosure in Teams UI	Privacy Office	Medium	Open

Approved by (name and role):

____________________________

Signature:

____________________________

Date:

____________________________

Audit trail: full machine-readable finding history, model IDs, timestamps, and evidence citations are stored in the Governance Fabric under initiative SAMPLE-03. Regenerate this pack at any time from the initiative page.