Key Takeaway
Data readiness assessments prevent the most common AI project failure mode: discovering data quality issues after investing months in model development.
When to Use This Template
Use this assessment before launching any AI project that depends on organizational data. It evaluates whether data assets are ready to support AI workloads across quality, accessibility, governance, infrastructure, and labeling dimensions. Running this assessment early prevents the painful discovery of data gaps months into a project. It is also valuable as a periodic health check for existing data assets that support production AI systems.
Assessment Dimensions
| Dimension | What It Measures | Key Questions | Red Flags |
|---|---|---|---|
| Data Quality | Completeness, accuracy, consistency, timeliness | What is the null rate for critical fields? How often are values incorrect? | No quality monitoring; unknown error rates |
| Data Accessibility | Cataloging, self-service access, documentation | Can engineers find and access data without filing tickets? | Data discovery requires tribal knowledge |
| Data Governance | Ownership, classification, consent, retention | Is every dataset assigned an owner? Is PII identified and tracked? | No data classification; unclear ownership |
| Data Infrastructure | Storage, processing, real-time availability | Can we process training data within our timeline? Is versioning in place? | No dataset versioning; manual processing |
| Labeling Maturity | Annotation workflows, QA, active learning | Do we have labeled data for our use case? What is label accuracy? | No labeling process; no quality metrics |
Scoring and Interpretation
Score each dimension from 1 (not ready) to 5 (production-ready). A dimension scoring below 3 should be considered a blocker for any AI project that depends on it. The overall data readiness score is the minimum score across all dimensions, not the average, because a single low-scoring dimension can block an entire project. Focus remediation efforts on the lowest-scoring dimension first.
data-readiness-scorecard.xlsx
XLSX · 142 KB
Data Readiness Scorecard with per-dimension scoring rubrics
Do not average data readiness scores across dimensions. A project with excellent data quality but no governance (score 1) will likely stall on compliance issues regardless of how good the data is. Your readiness is limited by your weakest dimension.
Version History
1.0.0 · 2026-03-01
- • Initial data readiness scorecard template