Idris Farouk runs a fourteen-person marketing operations department at a regional healthcare network. His boss came back from a conference fired up about AI and told him to "build an AI roadmap for the department." Idris's first instinct was to start buying tools and announcing initiatives. He caught himself. He had no honest picture of where his department actually stood: some of his analysts were quietly brilliant with AI, others had never opened a chatbot; his data lived in three systems that barely talked to each other; and he had no rule for who was allowed to use what. Building a roadmap on top of that fog would have been guesswork. So before planning where to go, Idris spent a week figuring out where his department really was. This lesson is the maturity assessment he ran, and you can run the same one on your own team.

What This Lesson Covers

A maturity assessment is an honest, structured snapshot of how ready your team or department is to get value from AI, scored across a few key dimensions so you can see exactly where you are strong and where you are stuck. You cannot set realistic goals or spend your budget wisely without it. This lesson gives you a five-dimension framework, a five-level scale to score each dimension, three ways to gather an honest picture (not just your own opinion), and a method for turning the scores into one concrete next step. We keep it at the altitude a manager actually owns: you are assessing your department, not the whole enterprise. We follow Idris as he scores his real team and picks his move.

The reason to start here rather than with a plan is simple. You cannot develop an effective strategy without understanding where you are. You cannot set realistic targets without knowing your starting point. And you cannot allocate resources sensibly without understanding what your organization is actually ready for. Most organizations are also uneven: one function is advanced while another is barely started, one team has strong data capability while another has almost none, leadership is excited while employees are skeptical. An assessment is what makes that unevenness visible instead of averaged away.

The Five Dimensions of Team AI Maturity

Maturity is never a single number. A department can be advanced in one area and stuck in another, and lumping it all into "we're a 2 out of 5" hides exactly the information you need. So you assess across five distinct dimensions:

• Skills: how many people genuinely know how to use AI well for their work, not just heard of it. Is it one enthusiast, half the team, or everyone?
• Tooling: do people have access to good AI tools, or are they making do with whatever free thing they found? Is there a sanctioned set of tools or a free-for-all?
• Process integration: is AI woven into how work actually gets done, with steps and templates, or is it a side activity a few people do when they remember to?
• Governance: are there clear, simple rules about what data can go into which tools and who decides what is approved? Or is it unmanaged?
• Data: is your team's information organized and accessible enough for AI to be useful on it, or is it scattered across systems and inboxes?

These five cover the things that actually determine whether AI helps your team. Notice that only one of them (tooling) is about technology. The rest are about people, habits, rules, and information, which is why buying tools alone never moves a team's maturity much.

Looking Closer at Each Dimension

The team-level list above is the working version. Behind it sits the fuller organizational framework, which asks the same questions in five slightly wider categories. It is worth knowing both, because the wider categories include two things the team-level list does not name at all, and those two are frequently where organizations actually get stuck. Idris scored his department against the working list, but he read it through the categories below to make sure he was not missing anything.

Technology and Infrastructure

Does the organization have the technology foundation for AI at all? This category covers the tooling and data dimensions together, and it includes four things: data infrastructure, meaning the systems for collecting, storing, and managing data; access to appropriate AI tools and platforms; integration capability, meaning the ability to fit AI into existing workflows and systems; and security and compliance systems that let AI be used safely and lawfully.

Strong technology maturity looks like modern data systems, access to quality tools, and the ability to integrate AI into existing workflows without significant friction. Weak technology maturity looks like fragmented data systems, limited access to tools, and integration that fights you at every step. Idris sat in between: good licensed tooling, but campaign and customer data spread across three systems, which is precisely the fragmentation this category is designed to expose.

Talent and Skills

Does the organization have the people required to implement and manage AI? Four sub-questions matter here. AI literacy: do most employees understand what AI is and what it can and cannot do? Technical skills: is there anyone who can implement, customize, and support AI systems? Change management skills: do managers know how to lead teams through AI-driven change? And business acumen: do teams understand how to connect AI to real business problems?

Strong talent maturity means widespread literacy, sufficient technical talent, managers who can actually lead change, and business teams who know how to drive adoption. Weak talent maturity means limited understanding, too little technical capability, managers unprepared for the change, and business teams who simply do not see AI as solving any problem they have. Idris's two skilled analysts sitting alongside twelve colleagues who had barely tried the tools is a textbook picture of the weak end.

Process and Governance

Does the organization have processes and governance for responsible, effective AI use? Five things belong here: clear decision-making processes, so it is obvious who decides whether a tool gets adopted; risk management processes for identifying and mitigating AI risks; data governance covering who controls what data is used and how it is protected; performance measurement, so you can tell whether AI is creating value; and escalation processes for what happens when AI causes a problem.

Strong process maturity is clear, lightweight governance that enables innovation while managing risk, where decisions happen quickly, risk gets identified and mitigated, and performance is measured. Weak process maturity fails in one of two opposite directions: no governance at all, which creates risk, or heavy governance, which creates friction and slows everything down. Both are failures. Idris's vague understanding that patient data should not be pasted anywhere, with no written rule and no named approver, was the first kind.

Culture and Mindset

Does the organization have a culture that embraces AI or one that resists it? Four signals tell you: acceptance of change, meaning whether people take up new ways of working or push back; risk tolerance, meaning whether people will experiment or want guarantees first; growth mindset, meaning whether people believe they can learn to work with AI or believe their skills are fixed; and trust in leadership, meaning whether people believe leadership is handling AI responsibly.

Strong culture maturity embraces change, takes intelligent risks, believes in learning, and trusts leadership, and people there see AI as an opportunity. Weak culture maturity resists change, avoids risk, treats skills as fixed, and distrusts leadership, and people there see AI as a threat. This dimension does not appear in the team-level list, and it is often the one that quietly decides whether anything you roll out survives contact with the team.

Strategic Alignment

Is the organization aligned on why AI matters and where it is heading? Four things again: a clear strategic vision people actually understand, investment prioritization people can see the logic of, defined success metrics, and leadership alignment, meaning leaders agree on direction rather than pulling in different ways.

Strong strategic alignment means clear vision, understood priorities, defined metrics, and leaders pulling the same direction. Weak strategic alignment means unclear vision, confused priorities, undefined metrics, or leaders who quietly disagree, and the visible symptom is different groups pursuing different AI goals. Idris's own situation started here: a boss energized by a conference, no stated vision, no metric, and an instruction to build a roadmap. Noticing that was itself part of the assessment.

The Five-Level Maturity Scale

For each dimension you place your team on a five-level scale. Use plain descriptions, not just numbers, so the rating is honest and hard to inflate:

• Level 1, Ad hoc: nothing intentional is happening. Any AI use is one person experimenting alone, invisible to everyone else.
• Level 2, Exploring: a few people are trying AI, getting occasional wins, but it is informal, uneven, and undocumented.
• Level 3, Operational: AI is a normal, expected part of how several real tasks get done, with at least some shared tools and guidance.
• Level 4, Systematic: AI is built into core workflows with templates, training, clear rules, and measurement; it is how the team works, not an extra.
• Level 5, Transformative: the team continuously improves its AI use, pioneers new approaches, and treats AI fluency as a core capability.

A realistic expectation: most teams sit at Level 2 or 3, and that is completely fine as a starting point. The goal of the assessment is not to feel advanced; it is to see clearly so your next move is the right one. Be honest. A flattering assessment that says Level 4 when you are really Level 2 will send your budget to the wrong place.

The Same Scale in Organizational Language

You will meet this scale again under a more formal set of names, and it is useful to recognize them because that is the vocabulary senior leadership tends to use. The common framework runs Initial, Managed, Defined, Optimized, Innovative.

• Level 1, Initial: the organization has only just begun thinking about AI. Few tools are in use, no formal processes exist, most people are not AI-literate, and leadership's vision is unclear.
• Level 2, Managed: some AI tools are implemented, basic processes are emerging, some people have real skills, and leadership is beginning to articulate a vision.
• Level 3, Defined: there are multiple tools and a clear process for adopting them, many people have AI skills, governance is clear, and the leadership vision is well articulated.
• Level 4, Optimized: processes are mature, governance is strong, AI skills are widespread, and continuous improvement cycles are running. AI is embedded in how the organization operates.
• Level 5, Innovative: the organization is pioneering new AI capabilities, leadership is thinking years ahead, the culture embraces AI-driven innovation, and the organization leads its field.

The same realism applies at this altitude. Most organizations sit somewhere between Level 2 and Level 3, and very few are genuinely at Level 4 or 5. If your assessment puts you at Level 4, check it twice.

Getting an Honest Picture, Not Just Your Own

Your view alone is biased; managers usually rate their team more mature than it is. Idris used three lenses together. Self-assessment was his own honest first pass against the five dimensions and levels. Team input meant actually asking his analysts where they felt strong and where they struggled, which surfaced things he had missed, including that two people were already using AI heavily but nobody had shared it. A light external glance meant comparing notes with a peer manager in another department and reading how similar teams describe their AI workflows, just enough to tell whether his team was roughly typical, ahead, or behind. The three together gave him a picture he could trust far more than his gut.

One warning Idris took seriously: a maturity assessment can make people defensive if it feels like a report card. He framed it openly as "let's figure out where we are so we can improve," involved the team in scoring, and never used the results to single anyone out. The point is learning, not judgment.

How to Run Each of the Three Lenses

Each lens has a method, and the methods are simple enough to run in a week.

Self-assessment. For every dimension, answer four questions in writing: what is our current capability, how strong are we, what gaps exist, and what would improvement actually look like? Be honest, because the point is not to feel good but to understand reality. You can widen this by asking each sub-team the same two questions, where are we strong and where do we need to improve, and comparing their answers to yours.

Stakeholder interviews. Interview leaders, team members, and stakeholders about how they perceive the organization's AI maturity. You will often discover that leadership believes the organization is more mature than it actually is, which is itself a finding worth having. Ask each person the same six questions: what is our current AI capability, what is working well, what are the biggest barriers, what should we improve, what are you excited about, and what are you concerned about. These conversations consistently produce richer insight than self-assessment alone, and they are where Idris learned that two of his analysts had quietly built workflows nobody else knew existed.

External benchmarking. Where possible, understand how your maturity compares to peers and competitors. This is genuinely hard because maturity information is rarely public, but there are four practical routes: read case studies about how other organizations are using AI, talk with peers at industry events, engage consultants who hold benchmarking data, and study competitors' job postings to see what skills they are hiring for. The purpose is not a precise ranking. It is knowing whether you are average, ahead, or behind for your industry.

Summarize the current state. Then write the whole thing down in the plainest possible form, one line per dimension with its level and a short reason, followed by an overall characterization. A department that lands at Level 2 on technology, Level 2 on talent, Level 2 on process, Level 2 on culture, and Level 2 on strategic alignment can reasonably describe itself as Level 2 overall, Managed. The summary exists so the assessment can be shown to someone else without a briefing.

A Worked Example: Scoring Idris's Department

Here is how Idris's marketing operations department actually scored across the five dimensions, with the honest reasoning behind each rating:

• Skills: Level 2 (Exploring). Two analysts were genuinely skilled, but the other twelve ranged from curious to never-tried. Real strength existed; it just was not shared.
• Tooling: Level 3 (Operational). The company had licensed a solid business-tier AI assistant and most of the team had access, so tooling was actually his strongest dimension.
• Process integration: Level 1 (Ad hoc). AI was not part of any documented workflow. The two skilled analysts used it their own way; nothing was templated or repeatable. This was his weakest dimension.
• Governance: Level 2 (Exploring). There was a vague "don't paste patient data anywhere" understanding but no written rule, no approved-tool list, and no clear approver. Risky for a healthcare network.
• Data: Level 2 (Exploring). Campaign and customer data sat across three systems; usable with effort, but not clean or centralized.

Averaging would have produced a misleadingly tidy "Level 2." Idris deliberately did not average. The spread was the insight: he had decent tools (Level 3) sitting on top of almost no process (Level 1), with skills locked inside two people. In other words, the department owned good equipment that almost nobody used in any organized way. That specific imbalance, not an average score, told him what to do next.

Reading the Gaps: Where to Act First

Maturity is supposed to be uneven, so resist the urge to make every dimension equal. Instead, find the dimension that is most holding back the others, your binding constraint. Two questions guide the choice: which gap is blocking value right now, and which gap, if closed, would unlock the dimensions you already have?

For Idris the answer was clear. Buying more tools (raising an already-strong Level 3) would have been wasted money. His real bottleneck was process integration at Level 1: he had the tools and even had two people who knew how to use them well, but none of that was captured in a repeatable way the other twelve could follow. Closing the process gap would simultaneously lift skills (by spreading what the two experts already did) and put his good tooling to actual use. One well-chosen move, several dimensions improved.

You can buy tools. You cannot buy the habits, shared skills, and simple rules that turn tools into results. That is why the weakest of those is almost always where a manager should invest first.

Three Gap Patterns You Will Probably Recognize

Assessments tend to produce a small number of recurring shapes, and naming yours makes the diagnosis faster.

• Technology mature, talent immature. You have the tools but people do not know how to use them. This was Idris's department almost exactly, and it is the most common shape in organizations that started their AI effort with a purchase.
• Talent mature, process immature. People have real skills but the organization has no clear processes to channel them, so capability stays personal rather than becoming organizational.
• Strategy aligned, culture resistant. Leadership knows where it wants to go, and the people do not want to go there. No amount of vision documentation fixes this one.

Identify which of your gaps is holding you back most, and remember that for most organizations the binding constraint turns out to be talent or culture rather than technology. You can buy tools. You cannot buy people who know how to use them, and you certainly cannot buy culture change. Prioritize accordingly: if talent is your biggest gap, invest in talent; if culture is, invest in culture.

From Assessment to One Concrete Next Step

An assessment that gets filed away changes nothing; the most common failure is doing the scoring and then never acting on it. So Idris converted his finding into a single, concrete, time-boxed move rather than a sprawling transformation plan.

His next step: over the next quarter, turn his two AI-skilled analysts into the source of three documented, templated AI workflows for the department's most common tasks (campaign-brief drafting, performance-summary writing, and audience-segment research), then train the rest of the team on those templates in two short sessions. Alongside it, he paired a tiny governance fix that cost almost nothing: a one-page rule stating which data tiers were allowed in the approved assistant and that patient data was never permitted, closing his riskiest gap at the same time. He set a simple success check: by quarter's end, at least ten of the fourteen analysts using the three templates in their normal work. That single move targeted his weakest dimension, leveraged his strongest, and gave him a measurable result, which is exactly what an assessment is for.

What Your Maturity Level Implies for Strategy

Idris's move made sense for a department at Level 2. It would have been the wrong move at a different level, because different maturity levels call for genuinely different strategies. The assessment is supposed to determine the strategy, not decorate it.

At Level 1, focus on building awareness so people understand what AI actually is, running pilot projects small enough to show value quickly, investing in the technology foundation, especially data systems, identifying the early adopters who are already excited, and establishing basic governance in the form of simple processes. Do not attempt to leap straight to Level 4. You will fail, because the foundational capability is not there to stand on.

At Level 2, focus on scaling the pilots that worked, deepening talent through training and hiring, formalizing processes so governance becomes structured rather than implied, broadening adoption beyond the early adopters, and getting leadership and teams aligned on strategy. You have proven AI works in your context. The job now is to scale it, which is exactly the logic behind Idris turning two experts into three shared templates.

At Level 3, focus on continuous improvement of results from the tools you already have, building advanced capability so use moves from basic to sophisticated, integrating AI into core processes until it becomes invisible and simply part of how work happens, building competitive advantage by positioning AI as a differentiator, and preparing for Level 4 through culture and deeper strategic thinking. You have the basics. Now build excellence.

At Level 4, focus on innovation with emerging capabilities, industry leadership that positions the organization as a leading thinker, attracting top talent by building a reputation as an AI-forward organization, scaling across the enterprise so AI becomes a core organizational capability, and using that capability for strategic differentiation in the market. You are advanced. The job is to leverage the advantage.

Treat the Assessment as a Repeating Measure

A maturity assessment is not a one-time verdict; it is a baseline you will compare against later. The first run tells you where you stand and what to fix. Its real power shows up the second time, when you can see whether the move you made actually shifted a dimension. Idris planned to re-score his department the same way at the end of the next two quarters, using the identical five dimensions and five levels so the comparison was honest.

This repetition does three useful things. It tells you whether your chosen investment worked: if process integration climbs from Level 1 to Level 3 after the template rollout, the move paid off; if it barely moves, you misdiagnosed the constraint or executed it poorly, and you adjust. It keeps the next step honest, because a dimension you ignored may quietly become your new bottleneck once you fix the old one; Idris expected that once process improved, data (still at Level 2 and scattered across three systems) might become the thing holding everyone back. And it gives you a simple, credible story for your own boss: "two quarters ago we were here, we did this, and now we are here," which is far more persuasive than a list of activity. Run it light, run it the same way each time, and let the trend, not any single score, guide you.

Three Traps in Maturity Assessment

Assessment without action. Scoring the team, nodding, and filing the result is the most common waste. The assessment exists only to drive a decision; if no investment or change follows, you have learned nothing useful.

Chasing equal scores. Trying to drag every dimension to the same level treats maturity as a tidiness exercise. Some dimensions naturally lead and some lag, and some matter more than others; your job is to fix the constraining gap, not to achieve uniformity.

Turning it into a verdict. If the assessment feels like a judgment, people defend their scores instead of improving them, and the conversation becomes about why the assessment is wrong rather than about what to do next. Frame it as a shared learning tool, involve the team in both assessing and improving, and keep it about the next step, never about blame.

Four Terms Worth Being Precise About

• AI maturity: the degree to which an organization has developed the capability, processes, and culture to implement and leverage AI effectively.
• Capability: an organizational competency or strength within a specific dimension, whether technology, talent, process, culture, or strategy.
• Gap analysis: identifying the difference between your current state and the future state you want, which is what tells you where to improve first.
• Strategic alignment: agreement among leadership and teams on why AI matters and where the organization is heading with it.

Practice and Reflection

Run these against your own department rather than in the abstract.

• Assess your organization across the five dimensions of technology, talent, process, culture, and strategy. Rate each from Level 1 to Level 5, then name your strongest and weakest dimension explicitly.
• Based on that assessment, identify your biggest opportunity for improvement. What specifically would change if you improved that dimension?
• Design an improvement plan for your weakest dimension. What would you actually do, what investment would it require, over what timeline, and how would you measure whether it worked?
• Talk with three people from different levels and different functions and ask each of them to assess the organization's AI maturity. Compare their answers to yours. Where do they see it differently, and what does that difference tell you?

Then reflect on three questions. What is your organization's AI maturity level, and what makes you say that? Which dimension is your biggest gap, why is it the biggest, and what would address it? And if you could invest in exactly one area to improve maturity, where would you invest and what would the return be?

Why This Assessment Is Worth the Week

Maturity assessment is a tool for honest self-reflection. It shows you where you are and where you need to go. It stops you making unrealistic demands of your organization, and it stops you spending money on areas that did not need it. The managers who use it well are the ones who are honest about the current state and strategic about closing gaps, which is a less comfortable combination than it sounds.

Understanding where your organization stands is the foundation of everything else in AI leadership. You cannot move forward without knowing where you are, and you cannot invest intelligently without understanding your constraints and your opportunities. Run an honest assessment, let it inform the strategy, invest in closing the gap that binds. Do that for a few cycles and maturity rises, and the advantage you get from AI starts to compound, which is exactly what Idris was after when he chose to spend a week looking before he spent a quarter building.

Key Takeaways

• Assess before you plan. You cannot build a sensible AI roadmap for your team without an honest picture of where it stands today; guesswork sends budget to the wrong place.
• Score five dimensions, not one number. Skills, tooling, process integration, governance, and data each tell you something different; an average hides the imbalance that actually guides your action.
• Use a five-level scale and be honest. Ad hoc to Transformative; most teams start at Level 2 or 3, and a flattering score is worse than a low one because it misdirects your investment.
• Get more than your own view. Combine self-assessment, real input from your team, and a light external comparison; managers consistently overrate their own team's maturity.
• Do not average; read the spread. The gap between your strongest and weakest dimensions is the real insight, as Idris's strong tooling over weak process showed.
• Fix the constraining gap first. Invest in the dimension that, once improved, unlocks the others; it is usually a people, habit, or rule gap, not a tooling one.
• Convert the assessment into one concrete, measurable move. A time-boxed next step with a success check is what turns a maturity score into actual progress instead of a filed document.
• Let the level choose the strategy. Awareness and pilots at Level 1, scaling and formalizing at Level 2, excellence and integration at Level 3, innovation and leadership at Level 4. Skipping levels does not work.
• Culture and strategic alignment count as dimensions too. They are the two that the team-level list does not name and the two that most often decide whether anything you roll out survives.

Frequently Asked Questions

Should I assess my whole company or just my team? Just your team or department, the part you actually control. Enterprise-wide maturity is leadership's concern; you would have no real ability to act on it. Assessing your own department keeps the exercise honest and, more importantly, keeps every gap you find within your power to do something about.

Why not just average the dimension scores into one maturity level? Because the average destroys the most useful information you have. A team with strong tooling and weak process averages to the same "Level 2" as a perfectly balanced one, yet the two need completely different actions. The gap between your highest and lowest dimensions is precisely what tells you where to invest, so keep them separate.

My team is mostly at Level 2. Isn't that bad? No, it is a normal and honest starting point; most teams are at Level 2 or 3. Maturity is a direction of travel, not a grade. A clear-eyed Level 2 with a concrete next step will out-improve a team that flatters itself as Level 4 and therefore does nothing. The level matters far less than whether you act on it.

How do I keep the assessment from feeling like a performance review? Frame it from the start as "where are we, so we can improve together," and assess the team's systems and habits, not individuals. Involve people in the scoring, share the results openly, and immediately point at the next step rather than the gaps. When people see the assessment leads to support and better tools rather than blame, they engage with it honestly.