Skip to content

04Personas/ Data leadership

Chief Data Officer / CDAO

Owns what AI is allowed to know — and often the whole AI strategy.

You are accountable for

Classification, lineage, and permissions that survive contact with retrieval
Training and grounding data with documented rights
Gartner finds ~70% of CDAOs hold AI strategy — this framework is their map

What keeps you up at night

An assistant surfacing the over-shared folder nobody audited
PII discovered in prompts, logs, or a vendor's retention store
A tuning corpus with rights nobody can prove

Decisions you own

Set the data-class-to-AI-system access matrix and enforce it at index time
Mandate de-identification pipelines for sensitive domains
Make lineage capture automatic from source to model to output

Your layers of the stack

2Data GovernanceWhat is the AI allowed to know?3Model GovernanceWhich models do we trust — and how do we know?1Enterprise GovernanceWho is accountable for AI — and for which AI?

Where to go first

Layer 2 — Data Governance Risk: data leakage Example: Wells Fargo zero-PII

Questions you should be asking

Which corpora are indexed for AI, and when were their permissions last reviewed?
Can we trace any AI output back to sources, model version, and training data?
Where does our data physically go — inference, caching, logging — per vendor?