AI Readiness Assessment Framework

Overview

This framework guides a structured assessment of organizational AI readiness. It goes deeper than the AI Maturity Model blog post — this is the working document I use in actual client engagements.

An AI readiness assessment answers three questions:

1. What AI initiatives are achievable given your current state?
2. What investments are needed to execute more ambitious initiatives?
3. In what order should you make those investments?

Dimension 1: Data Readiness

Data inventory

Before any AI initiative, catalog your data assets:

What you’re assessing: Do you have the data that your proposed AI use cases require? Is it clean enough to use? Can you get to it?

Data quality dimensions

For each critical data source, assess:

• Completeness: What % of records have all required fields?
• Accuracy: Spot-check 50 records. What % have errors?
• Consistency: Is the same thing represented the same way across systems?
• Timeliness: How stale is the data? For real-time inference, this matters a lot.
• Lineage: Do you know where the data came from and how it was transformed?

Data infrastructure maturity

Dimension 2: Team Readiness

Skill inventory

Map your current team:

The most common gap: Companies have data analysts but no data engineers. Data analysts answer questions about the past. Data engineers build the infrastructure that makes AI possible.

Critical skills assessment

For each person on your data/AI team, assess:

• Can they write production code (not just notebooks)?
• Have they deployed a model to production before?
• Can they monitor a model in production and detect degradation?
• Do they understand the business domain well enough to spot bad model outputs?

Dimension 3: Process Readiness

Experiment-to-production pipeline

Map your current process for taking an AI experiment to production:

1. Idea → Approved experiment: Who approves? What criteria?
2. Data access: How long does it take an engineer to get the data they need?
3. Development: Where does development happen? What’s the standard environment?
4. Evaluation: How is model quality evaluated? By whom?
5. Deployment: What does deploying a model look like? How long does it take?
6. Monitoring: How do you know when a model starts performing badly?
7. Retraining: When and how are models retrained?

Most companies that have “AI” in production have no answers for steps 6 and 7.

Decision-making process

• Who can authorize an AI initiative?
• Who decides whether a model is good enough to go to production?
• Who is accountable when an AI system makes a wrong decision?

Dimension 4: Infrastructure Readiness

Compute and deployment

Security and compliance

• Are there regulatory constraints on which data can be used for training?
• Is there a data retention policy that limits training data availability?
• Are there audit requirements for AI decisions?

Scoring and next steps

Rate each dimension on a 1–5 scale:

Score interpretation:

• 1–2: Focus on foundations before any AI initiative beyond simple automation
• 3: Ready for systematic AI in specific, well-scoped domains
• 4–5: Ready for ambitious AI initiatives

The typical finding: Most companies score 2–3 on data and team readiness, 1–2 on process and infrastructure. This is the gap that explains why AI pilots don’t scale.

What to do with this

Use this assessment to:

1. Identify the specific investments needed to advance AI maturity
2. Prioritize AI use cases that are achievable at your current maturity level
3. Build the business case for infrastructure investment
4. Set realistic expectations with leadership about AI timelines

The detailed Build vs Buy Framework covers the next decision you’ll face after the readiness assessment.