# Assessing AI maturity: where does your company stand?

Author: Marin Niehues, Niehues Consulting · Published: 2024-11-11 · Updated: 2026-11-16 · AI Strategy & Business Case
Source: https://www.niehuesconsulting.com/blog/ai-maturity-assessment/

You assess AI maturity by rating five dimensions on a five-stage scale against evidence and treating the weakest dimension as the constraint on the next step. Maturity describes how reliably a company produces results from AI applications in day-to-day operations. What counts is the number of applications that business units use every day, far more than the number of pilot projects.

**A company's AI maturity is capped by its weakest dimension, because every application needs all five dimensions at the same time to reach operation.**

## What is AI maturity, and what is it for?

AI maturity is the placement of a company on a staged scale that shows how reliably it brings AI applications into regular operation and measures their benefit. For the executive team the placement serves three purposes. It determines which initiatives are realistic next and justifies the order of investments. It also provides a baseline against which progress can be read after twelve months.

Without that baseline, the budget discussion rests on the experience of individual departments, and each department rates its own position generously. An evidence-based maturity rating replaces impressions with proof. How the assessment fits into the wider steering system for AI is set out in the article [AI management: what executive leadership must steer](/blog/ai-management-what-executives-must-steer/).

The assessment pays off as soon as more than two AI initiatives are running or a budget for the coming year has to be allocated. For a single first initiative, a check of the prerequisites is sufficient. Readiness asks whether the prerequisites for starting are in place. Maturity asks how well the company already delivers.

## Where do German companies stand on AI adoption?

One in eight companies in Germany with at least ten employees uses AI technologies, up from 11 percent two years earlier. This is shown by the official survey on ICT usage in enterprises. Large companies use AI considerably more often than small and medium-sized ones.

These figures measure usage. They say little about maturity, because a company already counts as a user when individual employees work with a tool. Anyone who wants to know where their own company stands therefore needs a model that maps the route into regular operation.

**AI use in German companies**
12 % | of companies with ten or more employees use AI technologies (survey year 2023)
11 % | was the share in survey year 2021

## Which AI maturity models exist?

Three types of model are in wide use: academic models, models from analysts and consultancies, and self-tests from software vendors. They differ in the number of stages and in how they weight the dimensions.

Academic models usually work with four to five stages, ranging from experimentation through pilots to AI embedded in the ways of working of the entire company. They describe which capabilities a company must build at each stage before the next becomes attainable.

Models from analysts and consultancies typically assess strategy, organization as well as culture alongside data and technology, and work with five stages. Vendor self-tests are quick to complete and often focus on the vendor's own platform. The model described here sits in the same line and adds a further yardstick, delivery capability: how reliably a company takes an idea all the way into regular operation.

## The five dimensions

The first dimension is strategy and steering. It asks whether the executive team has set AI goals linked to corporate goals, whether a budget exists and whether an owner is named. The second dimension covers use cases. The test is whether they were selected by value and feasibility, and whether a robust business case exists. The method is described in [Prioritizing AI use cases](/blog/ai-use-case-prioritisation/).

The third dimension is data and technology. It tests the availability and quality of data, access to the systems and the platform for operations. The fourth dimension is organization and enablement. It asks about roles and decision paths as well as the ability of the business units to work with AI. The fifth dimension is operations and impact. It concerns monitoring and quality assurance as well as the measurement of the benefit achieved.

**Self-test: one yes-or-no question per dimension**
- Strategy and steering: Is there a named owner for AI with a dedicated budget and target values?
- Use cases: Does at least one initiative have an approved business case with a measured baseline?
- Data and technology: Is the data for the most important use case accessible and of sufficient quality?
- Organization and enablement: Does a named person in the business unit accept the result and own its use?
- Operations and impact: Is the benefit of at least one application measured after launch and reported to the executive team?
Every answer of no marks a constraint. With two or more answers of no, the company sits at stage one or two.

## The five stages

At the first stage, Exploring, there are isolated points of contact. Employees use freely available tools, there is no common line and there are no owners. At the second stage, Experimenting, several pilots run in different areas. They arise from individual initiative, follow no prioritization and often end at the question of who takes over operations.

The third stage, Operationalising, separates experimenters from companies with measurable benefit. First applications move into regular operation, processes and responsibilities are defined, and the business units take ownership of use.

At the fourth stage, Scaling, the company repeats the route systematically. A shared platform and fixed roles make it possible to bring new applications into operation in a short time. The fifth stage is Steering and Renewing. At this stage AI is part of planning and controlling, and the executive team reads the benefit continuously from the portfolio.

**The five stages with typical evidence**
- Stage 1 Exploring: individual employees use freely available tools. Evidence: no list of initiatives exists.
- Stage 2 Experimenting: several pilots without prioritization. Evidence: a list of pilots, but no application in regular operation.
- Stage 3 Operationalising: first applications in regular operation. Evidence: adoption rate and measured benefit for at least one application.
- Stage 4 Scaling: a repeatable route into operation. Evidence: shared platform, fixed roles, several applications in operation.
- Stage 5 Steering and Renewing: AI embedded in planning and controlling. Evidence: a portfolio report with benefit per application in regular management reporting.

*Interactive tool: AI maturity self-assessment, assigns a company to a maturity level based on ten questions on strategy, use cases, data and technology, organization and operations and suggests next steps. Available at https://www.niehuesconsulting.com/blog/ai-maturity-assessment/*

## How to carry out the assessment

The assessment runs in three steps. In the first step you gather evidence: the list of all AI initiatives with their status, the existing business cases and the measured results. In the second step you hold short conversations with two or three people each from the executive team, IT and the business units, and compare their statements with the evidence. In the third step you assign a stage to each dimension, following the principle that a stage is only reached once it can be demonstrated in at least one productive application.

The most frequent distortion is overestimation. A pilot with a convincing demonstration is counted as an application although no business unit uses it. The evidence-based approach corrects this because it asks for use, ownership and impact.

**Routes to an assessment**
Criterion | Standard questionnaire | Internal self-assessment | With Niehues Consulting
Basis of the rating | Self-reporting | Perception of those involved | Evidence from initiatives and interviews
Focus on delivery | Low | Varies with the participants | Each dimension is tested against the route into regular operation
Comparability | Against average values | Against own expectations | Against the requirements of the company's own strategy
Result | Score | Sentiment overview | Prioritized measures with named owners

## How do you increase AI maturity?

You increase AI maturity by raising the weakest dimension by one stage while taking one application with a clear business case into regular operation. An improvement in a stronger dimension only increases delivery value once the constraint has been removed.

If use cases sit at stage three and data at stage two, the benefit only materializes once the data foundation of the most important use case has been built. If the organization lags behind technology and data, roles and decision paths are usually missing, and enabling the business units becomes the first initiative. If operations and impact is the lowest dimension, the work starts with measuring the benefit of the existing applications.

A twelve-month horizon suits the planning. In the first half-year the weakest dimension is raised by one stage and one use case is taken into regular operation. In the second half-year the next-weakest dimension and a second use case follow. Maturity is then assessed again and compared with the baseline.

## Niehues Consulting as a partner for maturity assessment

[Niehues Consulting](/) assesses AI maturity from the delivery perspective and supports companies from findings to regular operation. Marin Niehues builds organizations that deliver and brings AI into operations, in projects from VC-backed scale-ups to DAX 40 groups and in 20+ countries on six continents.

The basis is the Niehues Delivery Model with its phases Understand, Build and Deploy. The maturity assessment belongs to the Understand phase and answers its review question: what is the baseline, and how do we measure success? In the following phases, two further review questions matter most for maturity: which roles, decision paths and interfaces does operation require? Who owns operations from day one?

The result is a profile with prioritized applications, each with a business case, in a fixed sequence of measures. Prototypes follow that prove the benefit in the real workflow, and then deployment with clear processes and responsibilities. The entry point is a free [30-minute initial consultation](/#kontakt), in which the likely stage of each dimension and the available evidence are assessed.

**What to take away**
AI maturity is a delivery measure. Five dimensions with five stages and an evidence-based assessment produce a profile from which the executive team derives the sequence for the next twelve months. Set the priority at the weakest dimension.

**Sources**
- Federal Statistical Office of Germany (Destatis): About one in eight companies uses artificial intelligence, press release, 2023

## Frequently asked questions

### What is AI maturity?

AI maturity describes how far a company has progressed in bringing AI applications reliably into regular operation and measuring their benefit. It is usually expressed on a scale of four or five stages. Several dimensions such as strategy or operations are assessed separately.

### How can I test my company's AI maturity myself?

Answer the yes-or-no question from the self-test in this article for each of the five dimensions. Every answer of no shows a constraint, and with two or more answers of no the company sits at stage one or two. For a robust rating, check every answer of yes against evidence from a productive application.

### What is the difference between AI maturity and AI readiness?

AI readiness checks whether the prerequisites for starting with AI are in place. AI maturity measures how reliably a company already brings AI applications into operation. Readiness is the question before the first initiative, maturity the question from the second or third onwards.

### How long does an AI maturity assessment take?

With existing documentation and up to five business units involved, maturity is assessed in two to four weeks. Most of the time goes into gathering evidence and holding the conversations with the business units.

### How much does an AI maturity assessment cost?

The cost of an AI maturity assessment is driven by the number of units and interviews to be covered in two to four weeks. The market usually bills by consultant day.

### Who should be involved in the assessment?

The executive team and IT take part, as do the heads of the business units with current or planned applications. The business units supply the most important evidence, because that is where it becomes visible whether a result is used.

### How often should AI maturity be reassessed?

An interval of twelve months has proved suitable. During an ongoing implementation program, a shortened interim measurement after six months shows whether the weakest dimension has reached the planned stage.
