# The AI Maturity Model: Six Dimensions, Five Levels, and Where Your Company Stands

**By Dan Cumberland** · Published July 15, 2026 · Categories: AI Strategy

> Pacemark scores a company's AI maturity across six dimensions and five levels, from a few people using ChatGPT on the side to AI changing how the company wins work. Here's the whole model and how the scoring works.

Most companies describe their AI maturity by pointing at something they bought: a Copilot license for everyone, a pilot running in one team, a vendor demo that looked good in the room\. All of that tells you what's on the invoice, not what changed\. PwC's most recent global CEO survey found 56% of companies saw neither higher revenue nor lower costs from AI over the past year\. Maturity asks a different question: what can your people actually do with AI today, and what would it take to do more?

[Pacemark](https://pacemark.ai) is an AI maturity model built to answer that\. It scores a company across six dimensions and places each one on a five\-level scale, from a few people using ChatGPT off to the side, to AI changing how the company wins work, delivers it, and develops its people\. This page is the whole model: the six dimensions, the five levels, how the scoring works, and the research it came from\. If you want your own number instead of the general picture, [the assessment](https://pacemark.ai/signal/assessment/?utm_source=dcl-blog&utm_medium=blog-body&utm_content=ai-maturity-model) scores your company against it in about ten minutes\.

Most maturity models are a form a company fills out about itself\. I built Pacemark to score from what's actually happening in the work instead: the tools people reach for, the workflows that have changed, the policies that exist or don't\. That difference matters most at the company level\. A person can size up their own AI use, because they can see their own work\. Leaders rarely have that view across a whole workforce\. The company that says its customer\-success team uses AI everywhere often means two people use ChatGPT now and then\.

So the quick assessment on this page is a self\-estimate, enough to get you oriented\. The fuller company\-level picture comes from the discovery underneath the model: someone sitting with the teams, looking at the real work, and scoring the signal\.

## The five levels of AI maturity

A company's overall maturity lands on one of five levels\. Each level describes what you can see happening in the company\.

1. **Ad Hoc\.** People use AI on their own\. No strategy, no policy, no one steering\. Leadership may not even know it's happening\.
2. **Pilot\.** A few teams are trying things with permission\. There's interest and some leadership awareness, but no plan yet\.
3. **Integration\.** AI is built into specific workflows\. There's a written policy, and people are being trained on a real cadence\.
4. **Acceleration\.** The company wins work partly because of its AI capability\. Proprietary workflows exist, and ROI is tracked and reported\.
5. **Transformation\.** AI shapes how the company wins, delivers, and develops people\. New service lines exist that wouldn't without it\.

The levels came from watching companies move\. Across the interviews behind the model, the same progression kept showing up: a few people experimenting on their own, then pilots with permission, then written policy and training, then AI showing up in how the company wins work\. The levels put names on stages that were already there\.

## The six dimensions

Maturity isn't one number\. A company can have a sharp AI strategy and a workforce that's scared of it\. It can have every tool licensed and no governance at all\. The six dimensions came out of the research the same way the levels did: they're the areas that kept separating the companies making progress from the ones stuck at a pilot\. Pacemark scores each one separately so you see the real shape of where you stand, instead of one flattering average\.

**Strategy & Leadership\.** Executive sponsorship, a stated AI vision, a real budget line, and whether AI shows up in the strategic plan at all, or only in a slide someone made for the board\.

**People & Culture\.** The dimension most maturity models get wrong, and the one I built Pacemark to take seriously\. We measure two things: whether people feel safe experimenting, which we call Hearts, and whether they have the skill to do something real with AI, which we call Minds\. The two don't average against each other\. A workforce that's eager but untrained, or trained but too nervous to admit it touches AI, scores at the weaker of the two\. Most models file people as a footnote under "organization\." Pacemark weights People & Culture equally with strategy and technology, because the tools only give back what people actually use\. Take your people seriously, or the transformation stays on paper\.

**Data Readiness\.** Whether the data is clean, whether people can get to it, and whether there's a knowledge architecture: a place AI can go to find what the company actually knows, not just what's sitting in a folder somewhere\. That last piece is the one most companies haven't built yet\.

**Workflows & Operations\.** How deep AI runs inside the real work\. One experimental use case on the side, or a step written into the standard operating procedure that new hires get trained on\.

**Technology Infrastructure\.** The tools, whether they talk to each other, and whether anyone has thought about security\. A pile of licenses isn't infrastructure\.

**Governance & Ethics\.** Whether there's a written policy, whether someone owns AI risk by name, and whether anyone is actually watching what the tools do with client data\.

Your overall level is your **weakest** dimension\. If your governance is a Level 1, the company is a Level 1, even with Copilot on every desk\. We score it that way on purpose\. It stops a company from calling itself advanced because it spent a lot on software, and it points straight at the thing actually holding it back\.

## Where individuals fit: the five tiers

The People & Culture score comes mostly from where your people actually sit\. Everyone lands on one of five individual tiers\.

1. **Non\-User\.** Hasn't tried AI, or tried it and quit\.
2. **Curious\.** Uses AI weekly for search\-style tasks\. No saved patterns yet\.
3. **Practitioner\.** Uses AI weekly for at least one real task beyond search\. Has repeatable personal habits\.
4. **Builder\.** Builds processes and tools around AI\. Teaches teammates\.
5. **Architect\.** Builds things other people use\. Shapes where the company's AI goes\.

A company with most of its people at Tier 1 or 2 scores low on People & Culture no matter what the training deck says\. The distribution is the evidence\. The tiers feed that one dimension and nothing else, so the model stays six dimensions by five levels\.

## What this is built on

A maturity model is only as good as what's behind it\. Pacemark is built from **97 firsthand interviews**, covering **153 primary\-subject companies** and **300\-plus companies** in total, across **22 countries**, and cross\-referenced with **eight external benchmarks**: MIT CISR, PwC, BCG with Harvard and MIT, Bluebeam, Unanet, Stanford, McKinsey, and Accenture\.

MIT CISR's own research is a useful gut check on the top of that scale: across the companies they've studied, only 7% reach what they call Future\-Ready, where AI runs enterprise\-wide and shapes what the company sells\. Level 5 on this model is meant to be that rare\.

## How this compares to Gartner, MITRE, and the rest

There are plenty of AI maturity models to pick from\. MITRE publishes one, Gartner has one, KPMG has one, and universities have put out a stack of free ones\. If one of those fits, use it\. They're rigorous, and most cost nothing\.

Two things set this one apart\. The score comes from discovery, someone reading the signal in the actual work, where most models hand a company a form to grade itself\. And People & Culture is a full dimension weighted like the rest, scored mostly from where your workforce actually sits across the five tiers, which almost no published model does\. Companies come to Pacemark when they want the human side scored as seriously as the technical side, and a result they can defend to a board\. Other models count datapoints\. This one counts conversations\.

## Where your company stands

The fastest way to see where you stand is to take the assessment\. It walks the six dimensions, places you on the five levels, and gives you a current score, a picture of what the next level looks like in your specific work, and the path between the two\. It takes about ten minutes\.

[**See where your company scores on all six dimensions →**](https://pacemark.ai/signal/assessment/?utm_source=dcl-blog&utm_medium=blog-body&utm_content=ai-maturity-model)

If you want to go deeper on the ideas underneath the model, [what AI transformation actually means](https://dancumberlandlabs.com/blog/ai-transformation/) is the natural next read, and [AI literacy](https://dancumberlandlabs.com/blog/ai-literacy/) covers the tier ladder in more depth\. If you're closer to the starting line, the [AI readiness assessment](https://dancumberlandlabs.com/blog/ai-readiness-assessment/) is a lighter first step\.


---

Source: https://dancumberlandlabs.com/blog/ai-maturity-model/