The most common reason AI projects fail isn’t the technology chosen — it’s that no one checked whether the organization was ready for the technology. The AI Readiness Assessment is the tool that performs this check before the project begins.
The 5 readiness dimensions — each scored from 0 to 5.
🎯 What is an AI Readiness Assessment
An ARA is a structured assessment that measures the organization's ability to adopt AI and get results. It doesn't measure leadership's enthusiasm — it measures infrastructure, data, people, processes, and governance. It's the instrument that turns assumptions about readiness into objective data.
🔑 Why do this before any project
- •Prevents doomed projects — low readiness + an ambitious project = guaranteed failure.
- •Directs the first investment — the priority gap is the first step in the roadmap.
- •Calibrates expectations — the client understands why "6 months of preparation before AI" isn’t a waste of time.
💡 Practical tip
Use the radar as a conversation tool in the first meeting with the client — not as a spreadsheet sent by email. Filling it out together builds alignment and reduces resistance to difficult recommendations.
💾 Dimension 1 — Data
It’s the dimension most critical and most underestimated. Companies that have never structured data for analysis are rarely ready for AI. Assessing quantity, quality, and access determines whether AI has the fuel it needs to work.
✓ Data ready for AI
- ✓History > 12 months with complete fields
- ✓Consistent format across systems
- ✓Access via API or reliable export
- ✓Representative sample without obvious bias
✗ Data not ready
- ✗Data in manual spreadsheets with no standardization
- ✗Critical fields are empty (>20% missing)
- ✗Legacy system with no export
- ✗Data from only one customer segment
💡 When the recommendation is: “fix the data first”
If the data dimension scores below 3, the honest recommendation is to pause the AI discussion and prioritize data quality. That recommendation may seem counterintuitive, but it’s exactly what builds trust—and protects the future project from failure.
⚙️ Dimension 2 — Technology and infrastructure
Infrastructure defines what is technically possible. It’s not just about having “cloud” — it’s about connectivity between systems, processing capacity, and whether the current stack supports the AI components that make sense.
Score 0-2: Basic or legacy infrastructure
Systems without an API, old on-premise systems, isolated data. AI will require infrastructure modernization first—which significantly affects timelines and budgets.
Score 3: Partially connected
Some systems have APIs, and some data is accessible. You can start with specific use cases while modernization moves forward in parallel.
Score 4-5: Cloud-ready with APIs
Modern stack, documented APIs, integration capabilities. Ready to receive AI components with no infrastructure blockers.
👥 Dimension 3 — People and culture
The most often overlooked in AI projects. The best technical solution fails to gain adoption if the team doesn’t understand what they’re getting, distrusts the technology, or has no one internally to support the project after the consultant leaves.
🧑🤝🧑 What to assess in people and culture
- •AI literacy: does the team understand the basics of how AI works and its limits?
- •Internal champions: is there someone who will lead adoption after you leave?
- •Change history: has the organization successfully adopted new technologies before?
- •Engaged leadership: Does C-level or management actively sponsor the initiative?
basics of models and limitations
who supports it after the consultant
adoption of past changes
visible and vocal leadership
🔄 Dimensions 4 and 5 — Processes and governance
Processes and governance are the operational backbone. Without documented processes, AI enters chaos and amplifies it. Without governance, the organization operates on the hope that nothing goes wrong—and when something does, the damage is greater.
🔄 What to assess in processes
- • Documented critical processes
- • Clear inputs and outputs for each workflow
- • Controlled variability (same result under the same conditions)
- • Ability to integrate an AI step without breaking the workflow
🏛️ What to assess in governance
- • Is there an AI use policy?
- • Is someone assigned responsibility (CAIO or equivalent)?
- • Have AI risks been mapped?
- • Is there a defined process for reviewing AI outputs?
📌 Important note
A poorly defined process is chaotic when handled by people. The same process with AI is chaotic and fast—amplifying errors at automation speed. Documenting and stabilizing the process before automating isn’t bureaucracy; it’s a prerequisite.
📊 How to Score and Present the Radar
The 0-to-5 score for each dimension is deliberately simple. Accuracy isn’t the goal — the conversation is. The radar makes gaps visible and creates alignment on what needs to be resolved before any implementation.
📐 Scoring scale by dimension
ready for ambitious projects
ready with scope constraints
prepare before implementing
pause — fix the foundation first
🎒 Module summary
Next module:
3.2 — AI maturity models: a mirror for calibrating expectations and sequencing the roadmap