On this page
Key takeaways
- AI readiness is an organisational question as much as a technical one: strategy, sponsorship and change management matter as much as data and infrastructure.
- Score each of the ten dimensions from 1 to 5 with honest evidence, not aspiration.
- Your lowest scores, not your average, determine what will stall your first AI project.
- A score of 20–30 means start with foundations and one narrow pilot; 31–40 means pilot and scale selectively; above 40 means build a portfolio.
- Repeat the assessment every six to twelve months to track progress and reprioritise investment.
An AI readiness assessment tells you whether your organisation can turn AI investment into measurable business value, and what to fix first if it cannot. The quickest way to run one is to score ten dimensions from 1 to 5: strategy, executive sponsorship, data, infrastructure, security, skills, process, governance, budget and change management. Your total shows where you stand overall, and your lowest scores show what will stall your first project.
Most leadership teams skip this step. They pick a use case, hire a vendor or spin up a pilot, and discover three months in that the data lives in spreadsheets, nobody owns the outcome and legal has not approved the tool. A two-hour honest assessment upfront saves that quarter.
Why should leadership assess AI readiness before investing?
AI projects rarely fail because the model is bad. They fail because the organisation around the model is not ready: unclear goals, missing data, no business owner, or teams that quietly refuse to change how they work.
An assessment does three things for you. It replaces opinion with evidence, so the CTO, CFO and COO are debating the same facts. It surfaces blockers early, while they are cheap to fix. And it sets a baseline, so you can show the board progress in six months rather than just activity.
It also protects you from both extremes. Some companies overestimate readiness because they have a data warehouse. Others underestimate it and wait for perfection. A structured score keeps the conversation grounded.
The 10-point AI readiness checklist
Work through each item as a leadership team. For each, write down the evidence, then agree on a score using the rubric in the next section.
1. Strategy
Do you have a clear view of why you are adopting AI, tied to business outcomes? Look for a short list of priority problems, each linked to a KPI such as margin, cycle time, customer retention or risk exposure. "We need an AI strategy because competitors have one" scores low. "We want to cut order-to-cash time by a meaningful amount and AI-assisted document processing is our first lever" scores high.
2. Executive sponsorship
Is there a named senior leader who owns AI outcomes, has budget authority and will remove obstacles? Sponsorship means showing up to steering meetings and making trade-offs, not just approving a slide. Without it, AI initiatives get deprioritised the moment quarterly targets tighten.
3. Data
Is the data your priority use cases need available, accessible, reasonably accurate and owned? You do not need perfect company-wide data. You do need to know where the relevant data lives, who owns it, how clean it is, and whether it can be joined across systems. Ask a simple test question: could someone pull two years of the relevant history within a week?
4. Infrastructure
Can your technology estate support AI in production? That means cloud or on-premise capacity, APIs into core systems such as ERP and CRM, and the ability to deploy, monitor and update models or AI services. Legacy systems with no integration layer are a common hidden blocker.
5. Security and privacy
Do you have clear rules for what data can be sent to external AI services, and technical controls to enforce them? Consider customer data, employee data, intellectual property and regulated information. If staff are already pasting contracts into public chatbots, your score is low regardless of policy documents.
6. Skills
Do you have the people to specify, build, buy, evaluate and run AI solutions? This includes technical skills (data engineering, AI engineering) and business skills: product owners who can define requirements and managers who can interpret AI output. Many mid-size companies score low here and fill the gap with a partner while upskilling internally.
7. Process
Are the processes you want to improve documented, stable and measured? AI amplifies a process; it does not fix a broken one. If three regional teams handle the same task three different ways, standardise first or the AI will struggle.
8. Governance
Is there a lightweight framework for approving AI use cases, managing risk, assigning accountability and reviewing performance over time? Good governance covers model accuracy, bias, explainability where it matters, human oversight, and regulatory obligations in the markets you serve. It should enable fast, safe decisions, not add months of committee reviews.
9. Budget
Is there dedicated, realistic funding, including run costs after launch? Budgets often cover a pilot but not integration, change management, monitoring and scaling. Look for a funding model that releases money in stages as value is proven.
10. Change management
Are your people prepared to adopt new ways of working? This covers communication, training, incentives and addressing fears about job impact. The best technical solution delivers nothing if the sales team keeps using its old spreadsheet.
The scoring rubric: how to rate each dimension from 1 to 5
Use this table to keep scores consistent. Score what is true today, backed by evidence, not what is planned.
| Score | Label | What it looks like |
|---|---|---|
| 1 | Absent | Nothing in place. The topic has not been discussed seriously at leadership level. |
| 2 | Emerging | Informal awareness and some individual efforts, but no ownership, documentation or funding. |
| 3 | Developing | Defined in at least one area or function, with an owner. Gaps are known but not yet closed. |
| 4 | Established | Consistent across the relevant functions, documented, funded and used in real decisions. |
| 5 | Leading | Mature, measured and continuously improved. Could be used as an example for peers. |
Example: scoring the data dimension
To make the rubric concrete, here is how data might score in a mid-size manufacturer:
- 1: Production and sales data sit in separate spreadsheets maintained by individuals.
- 2: An ERP holds most transactional data, but exports are manual and definitions differ by plant.
- 3: A central data store exists for sales and inventory, with a named owner, but quality issues are common.
- 4: Core data is integrated, documented and quality-checked; teams trust the dashboards.
- 5: Data is governed as a product, with quality metrics, lineage and self-service access for approved users.
Run the same exercise for each dimension. Disagreements between leaders are useful; they usually reveal that one function has a very different experience from another.
What should you do with your AI readiness score?
Add up the ten scores for a total between 10 and 50. Then look at the total and the lowest individual scores.
| Total score | Readiness level | Recommended next move |
|---|---|---|
| 10–19 | Early | Focus on foundations: agree strategy, appoint a sponsor, set basic security rules. Run AI literacy sessions for leadership. Avoid large builds. |
| 20–30 | Foundational | Pick one narrow, high-value use case with available data. Fix the specific gaps that use case exposes. Consider a partner for skills. |
| 31–40 | Ready to pilot and scale | Run two or three pilots with clear KPIs. Formalise governance and a funding model that scales winners. Start building internal capability. |
| 41–50 | Advanced | Manage AI as a portfolio across functions. Invest in platforms, reusable components and continuous measurement. |
Why your lowest score matters most
A total of 35 sounds healthy. But if that 35 includes a 1 on executive sponsorship or security, you have a blocker that will stop any project regardless of how strong your data is.
Apply a simple rule: any dimension scoring 1 or 2 must have an action owner and a 90-day plan before a pilot goes live. Scores of 3 can be improved in parallel with a pilot.
A 5-step action plan after the assessment
- Record the scores and evidence in a single page the whole leadership team signs off.
- Name the two or three weakest dimensions and assign an owner to each.
- Choose one use case that is valuable and realistic given your current scores, not your hoped-for ones.
- Set 90-day targets for both the pilot (business KPI) and the readiness gaps (for example, "security policy approved and communicated").
- Reassess in six to twelve months and share the movement with the board.
Common mistakes when assessing AI readiness
A few patterns undermine the value of the exercise.
Letting IT run it alone
Readiness is a cross-functional question. If only technology leaders score it, data and infrastructure get detailed attention while sponsorship, process and change management are glossed over. Include operations, finance and at least one frontline business unit.
Scoring aspirations instead of reality
"We are about to launch a data platform" is not a 4. Score what is live and used today. Optimistic scoring leads to overconfident plans and missed deadlines.
Treating readiness as a gate rather than a guide
Some organisations use a low score as a reason to do nothing. Readiness improves fastest when you work on it alongside a focused pilot, because the pilot shows exactly which gaps matter. Use the score to choose the right first step, not to postpone indefinitely.
Assessing in the abstract
Readiness only means something relative to a use case. A retailer may be highly ready for AI-assisted customer support, because ticket history is clean and the process is well defined, yet poorly ready for dynamic pricing, where data is fragmented and the commercial risk is higher. When you score, keep your two or three candidate use cases in view and note where a dimension is strong for one and weak for another. That turns a generic score into a practical shortlist of where to start.
Assessing once and forgetting
Readiness changes as teams, systems and regulations change. A repeat assessment every six to twelve months keeps investment aligned with reality and gives the board a clear progress story.
How P26 helps
If you want an independent, evidence-based view, our AI Readiness Audit goes beyond self-scoring with stakeholder interviews, data and systems reviews, and a prioritised roadmap tied to your KPIs. For leadership teams ready to act on the results, our AI strategy consulting helps you choose the right first use cases, set up practical governance and build a plan the board can back. We work with CXOs and senior managers who want clarity on where AI genuinely matters in their product, operations, sales and processes, without the hype.
Want to know where your organisation really stands before you commit budget? You can book a call, and we will walk through the checklist with you.
Frequently asked questions
What is an AI readiness assessment?
An AI readiness assessment is a structured review of whether your organisation has the strategy, leadership, data, technology, skills, governance and culture to deliver value from AI. It produces a score per dimension and a prioritised list of gaps to close before or alongside your first AI initiatives.
How long does an AI readiness assessment take?
A leadership self-assessment using a checklist like this one can be done in a two to three hour workshop. A deeper assessment with data audits, system reviews and stakeholder interviews typically takes two to four weeks.
Do we need perfect data before starting with AI?
No. You need data that is good enough for one specific, valuable use case, and a plan to improve it. Waiting for company-wide perfect data is one of the most common reasons AI programmes never start.
Who should be involved in an AI readiness assessment?
Include the CEO or a business-unit head as sponsor, plus leaders from technology, operations, finance, and one or two functions that would use AI first. Having IT answer alone produces an overly technical and usually optimistic picture.
- #ai readiness
- #ai strategy
- #checklist
- #ai governance
- #change management
