Carnegie Mellon and Accenture Built a Free AI Maturity Model. Here Is What It Tells Leaders.

The headline is not the model. The headline is the number the model exists to fix.

SEI points to two figures to explain why it built this. 95% of organizations are realizing no returns on their AI investments. Only 8% are scaling AI at the enterprise level. The first number comes from the State of AI in Business 2025 Report. The second comes from Accenture's own Front-Runners' Guide to Scaling AI.

Read those two numbers together and the picture is stark. Heavy investment across the board. Almost no return to show for it. And a tiny share that has figured out how to scale.

If you feel behind on AI, this is the most useful thing to understand right now: you are not behind the way you think you are. The companies ahead of you are not ahead because they bought more tools. They are ahead because they built the practices underneath the tools.

What the model actually measures

The AI Adoption Maturity Model is built to help organizations move past experimentation and scale AI with measurable, repeatable outcomes. It helps a leadership team see capability shortfalls and build a roadmap for responsible adoption.

It does this two ways: a set of levels and a set of dimensions.

Five levels of maturity

The model defines five stages an organization moves through:

• Exploratory AI
• Implemented AI
• Aligned AI
• Scaled AI
• Future-Ready AI

The names tell the story on their own. You start by exploring. You put some things into production. You get the organization aligned. You scale what works. And then you build for what comes next.

The point is not which label you wear today. The point is that maturity is a climb, and you can measure where you stand before you decide where to spend.

Eight dimensions of readiness

The model assesses readiness across eight areas:

• Organizational strategy
• Workforce and culture
• Workflow re-engineering
• Risk and governance
• Data
• Engineering
• Operations
• Ecosystem

Notice what is on that list. Strategy, people, process, governance. Only one of the eight is the engineering work most people picture when they hear "AI adoption." The other seven are the parts a leader owns directly.

This is where the model earns its credibility. SEI has spent decades giving industries a structured way to measure readiness and reduce risk. The team built version 1.0 by reviewing more than 100 existing AI maturity models, analyzing three dozen of them in depth, and surveying nearly 600 practitioners. Then they validated it in pilots with Fortune 500 organizations.

Bosch Global Software Technologies is named as a real-world user. Srinivasulu Nasam, who heads enterprise AI transformation there, described the assessment as more than a point-in-time evaluation. He said it gave the company "a structured, actionable understanding of where we are succeeding, where more attention may be needed, and how to prioritize future investments for maximum ROI."

That is the use case in one sentence. Know where you stand. Know where to put the next dollar.

The disconnect a leader can close this week

Accenture's research surfaces a pair of numbers that sit right next to each other and tell on the whole industry.

86% of C-suite leaders plan to increase AI spending in 2026. Only 21% of organizations are redesigning end-to-end processes with AI at the core.

Sit with that distance for a second. Almost nine in ten leaders are about to spend more. Roughly one in five has changed how the work actually gets done. Accenture also reports that nearly half of executives say AI has delivered little impact on profit so far.

More spending on the same workflow, and no profit to show for it. The missing step is sequence. The workflow has to be redesigned before more spending can pay back.

AI amplifies the workflow it inherits. Point a capable system at a clean, well-designed process and it makes that process faster and cheaper. Point the same system at a broken process and it makes the mess faster.

Here is the operating principle worth carrying out of this story. The work that pays back happens before deployment: designing the system the agent will run inside.

Ipek Ozkaya, a technical director at SEI, said it plainly. "Our industry often assumes discipline can be automated away. But sustainable AI success still depends on disciplined engineering, governance, and operational practices."

You cannot buy your way out of that with a bigger software budget. You build it.

Why this maps to the 7 Levels of AI Proficiency

We have been carrying a version of this message for a while, so seeing it arrive from a Fortune-500-piloted, university-backed source is worth pausing on.

The 7 Levels of AI Proficiency starts from the same premise the SEI model does. Capability is measurable. It is a climb across distinct skills, not a single switch you flip. The volume of AI you deployed is the wrong scoreboard. The right scoreboard is how capable your people and your processes have become.

The eight dimensions in the SEI model give a leader a concrete checklist. Strategy, workforce and culture, workflow re-engineering, governance, data, engineering, operations, ecosystem. Run an honest read across those eight and you will find your real position quickly. Expect strength in one or two and thinner ground in the rest. That is normal. It is also the roadmap.

Manish Sharma, Accenture's Chief Strategy and Services Officer, drew the line between models that stop at slogans and models that hold up under pressure. "Many AI maturity models in the market now focus on high-level strategy without considering the engineering rigor that organizations need to actually scale," he said. The framework, he added, is "designed to meet organizations where they are across eight critical dimensions of AI readiness."

Meet organizations where they are. That is the invitation in both the SEI model and the 7 Levels of AI Proficiency. You do not need to be at the top to start. You need to know your floor.

What to do with this

You do not need a consulting engagement to get value from this release. Three steps, this week.

Download the model. It is free from the SEI Digital Library. Read the five levels and the eight dimensions. Spend twenty minutes, not twenty hours.

Do an honest read across the eight dimensions. Pick one number per dimension on a simple low-to-high scale. Be honest about governance and data, the two most people overstate. The exercise is only useful if it stings a little.

Pick the workflow before you pick the tool. Choose one core process your team runs every week. Map how it works today. Decide what a better version looks like. Only then ask where AI fits inside it. That order is what separates the 21% redesigning their work from the 86% adding to a budget.

The companies in that 8% scaling at the enterprise level did not get there by accident. They built strategy, people, process, and governance to a level where AI had something good to amplify. You can build the same thing. The first step is knowing exactly where you stand today.

Related reading: Level 5: Captain.

Sources

1. SEI and Accenture release AI Adoption Maturity Model to help organizations scale AI with predictable outcomes
2. Accenture and the Carnegie Mellon University Software Engineering Institute launch AI Adoption Maturity Model
3. SEI Digital Library: AI Adoption Maturity Model
4. State of AI in Business 2025 Report
5. Accenture: Front-Runners' Guide to Scaling AI

Harrison Painter

Executive AI Advisor. Founder, LaunchReady.ai and AI Law Tracker.

Harrison is an Indiana AI Advisor who helps business owners and executives get their time back by building AI systems that run the work for them. Nearly 20 years in business and author of You Have Already Been Replaced by AI. Creator of The 7 Levels of AI Proficiency.

Connect on LinkedIn