The question is not how much AI you use, but whether you can sustain it
A company can have licences, several pilots and employees using assistants and still not be ready to scale. Artificial intelligence maturity measures whether the organisation can turn those experiments into repeatable, secure and measurable results. It is a diagnosis of business capability, not an inventory of tools.
It is worth carrying out before approving new projects, when preparing an annual budget or when pilots fail to reach production. It also helps organise scattered initiatives and connect the AI transformation with real business objectives.
A mature company is not the one that buys the most tools. It is the one that knows what problem it is solving, who is accountable for the result and what it will do if the system fails.
The five dimensions you should assess
- Strategy and leadership: there is a clear ambition, priorities and an owner with decision-making authority.
- Data and technology: sources are accessible, have sufficient quality and can be integrated without creating a disproportionate architecture.
- People and culture: teams understand what to use AI for, receive training and participate in redesigning the way work is done.
- Governance, risk and compliance: there is an inventory of systems, risk levels, controls, oversight and traceability.
- Value and measurement: every initiative has a baseline, business metrics, total cost and criteria for continuing, correcting or stopping.
Score each dimension from 1 to 5 and require evidence. A corporate presentation does not demonstrate leadership; a policy without owners does not demonstrate governance; having data does not demonstrate data quality. Note which document, metric or process supports each score.
How to interpret the result without hiding bottlenecks
Avoid compensating for a critical weakness with strengths in other areas. An excellent technical team does not make up for the absence of an objective; committed leadership does not fix fragmented data. In addition to the average, look at the lowest score: it usually points to the real limit on growth.
Classify the diagnosis into three levels: exploration, when isolated tests predominate; operationalisation, when some use cases work with owners and metrics; and scaling, when shared capabilities exist to launch, monitor and govern several systems. The level must be assigned based on evidence, not aspiration.
From diagnosis to a 90-day plan
- Select the two gaps that block the most initiatives, not the ones that are easiest to solve.
- Define one observable action for each gap: a data catalogue, usage policy, role-based training or dashboard.
- Assign an owner, budget, date and completion evidence.
- Choose one or two use cases that can demonstrate the new way of working.
- Repeat the assessment at the end of the quarter and record what improved.
If the organisation is at an initial level, prioritise fundamentals and one small-scope use case. If it is already operationalising AI, create reusable components: risk assessment, contracts, observability and a deployment methodology. If it is scaling, focus on the portfolio, costs, interoperability and learning across teams.
Signs of false maturity
- Many pilots, but none has a business owner.
- Success is measured by usage or the number of prompts, not by results.
- Each department buys tools and manages data on its own.
- Security reviews projects at the end, when changing the design is already expensive.
- Employees use AI, but do not know what information they may enter.
- There is no plan for retiring models or providers.
AI governance must grow proportionally. An SME does not need to copy a multinational’s structure, but it does need explicit decisions, owners and records that make it possible to demonstrate how it works.
What should remain when you finish
The useful deliverable is a capability map with evidence, prioritised gaps, dependencies and a quarterly plan. Add an inventory of initiatives and link it to risks, required data, affected users and metrics. This turns the diagnosis from a snapshot into an investment tool.
AI maturity assessment template
Create a table with one row per capability and these columns: observed situation, evidence, score, risk of inaction, dependency, action, owner and date. Add a confidence column to distinguish an assessment based on data from a perception that has not yet been validated.
In strategy, look for an approved portfolio and budget; in data, quality reports and owners; in people, usage by role and assessments; in governance, an inventory and recorded decisions; and in value, baselines and outcomes. Interview leadership, business, technology, security and users: differences between answers reveal alignment problems.
Frequently asked questions about AI maturity
Is a full technical audit necessary?
Not for the first assessment. Start with enough evidence to identify blockers. Go deeper technically into the data, integrations or systems that affect the prioritised use cases.
What score allows us to start?
There is no universal number. A reversible, low-risk use case can start with limited maturity; a sensitive decision requires stronger capabilities. The score guides scope; it does not grant automatic permission.
Who should own the assessment?
A business owner with cross-functional support. Technology can coordinate data and architecture, but maturity includes adoption, value, risk and strategy, so it should not be confined to IT.
What we are learning at Impulsa3 by measuring AI maturity
In my experience at Impulsa3, I3OS has forced us to measure maturity by the real ability to work with AI, not by the number of licences or activated tools. We have had to organise strategy, data, people, technology, governance and measurement so that an SEO pilot could become a shared system for the whole company. That is the learning we apply when assessing what can be sustained and what still needs method, owners or evidence.
If you need to assess your company’s artificial intelligence maturity and turn the diagnosis into an actionable plan, Impulsa3 can support you throughout the process with a practical approach.
