AI Data Readiness: Check Sources Before Building

Define the decision and data need

State the workflow, question, users, output, consequence, and minimum information required.

The AI and machine-learning services route provides an internal path when data readiness needs technical assessment.

Establish source authority and lineage

Name the source owner, system of record, transformations, permitted use, access role, and point at which data can no longer be trusted.

Check quality, freshness, and structure

Review missing values, duplicates, definitions, dates, formats, relationships, changes, and known exceptions.

Control access and permitted use

Define who may access the data, what the workflow may retain, sensitive fields, and what must be removed or masked.

Turn gaps into a readiness decision

Record what is ready, what needs remediation, which risks are accepted, who owns the gap, and what test would change the decision.

Questions teams ask about AI data

More data does not always improve a workflow. Relevance, authority, quality, access, and context matter.