AI

What is a digital maturity assessment?

November 20, 20266 min read
Operations and data leads mapping capabilities at a whiteboard, for a post on what a digital maturity assessment is

A digital maturity assessment scores how far an organization has come in using digital technology, data and automation, across dimensions such as strategy, data, people and AI, against defined stages. Used well, it ends in one workflow with a number to move. Future Works ships that workflow in a 12-week cycle, with part of its fee staked on Finance's verdict.

Most assessments share the same skeleton: a set of dimensions, a scale for each, a score, and a comparison with peers or with the organization's own earlier result. Where they differ is in what they do with the score afterwards, and that part decides whether the exercise was worth the time.

What is a digital maturity assessment?

It is a structured way of measuring how capable an organization is of running its work on data, software and automation today, and where the gaps are. The answers usually come from a questionnaire, interviews with leaders and operators, and a look at the systems and data themselves. Each area gets a score on a maturity scale, from early and manual at one end to integrated and data-driven at the other.

A clear public example is the European Commission's Digital Maturity Assessment tool for small and mid-sized companies, built by its Joint Research Centre and used by the European Digital Innovation Hubs. The tool is a scored questionnaire, filled in before a hub starts working with a company, and the hubs then measure the same company again one year later and two years after that.

What does a digital maturity assessment measure?

The dimensions vary by model, and most cover the same ground. The EU tool uses six: digital business strategy, digital readiness, human-centric digitalisation, data management and connectedness, automation and artificial intelligence, and green digitalisation. Other models group them differently.

AI now has maturity models of its own. MIT's Center for Information Systems Research published an Enterprise AI Maturity Model in December 2024, based on a survey of 721 companies, with four stages: experiment and prepare (28% of the companies), build pilots and capabilities (34%), develop AI ways of working (31%) and become AI future ready (7%). Companies in the first two stages had financial performance below their industry average, and companies in stages three and four were well above it. Its authors suggest using the model to assess capabilities across dimensions such as processes, technology and organizational culture.

If you want to see how organizations in your industry tend to score, our post on digital maturity benchmarks by industry covers the stages and the comparisons.

How is a digital maturity assessment done?

An assessment usually runs in three steps. First, the scope: the whole enterprise, one business unit, or one function. Second, the evidence: a questionnaire for a broad read, interviews for the reasons behind the answers, and a review of the actual systems and data, which is where self-reported scores get checked. Third, the scoring: each dimension placed on the scale and compared with peers, an earlier result or a target.

Who takes part matters as much as the method. MIT CISR's advice for its AI model is to bring a team of senior technical and data leaders together to agree which stage the enterprise is in today and what it is aiming for. I would add the people who run the operation day to day, because they know which spreadsheet the whole process depends on, and a survey answered at headquarters rarely finds it.

What should a digital maturity assessment produce?

A score on its own changes very little. The useful output is a short list of gaps ranked by what closing them is worth, and, at the top of that list, one workflow where the gap costs real money and a number Finance already tracks would move if it closed.

As I wrote in August, the useful question for AI is which data a system can safely touch today, asked domain by domain: someone named who owns the dataset and stands behind it, one definition of each field across the systems that use it, and a trail showing where a number came from. Many organizations have governance policies on paper and still fail that test, the gap I called governance espoused versus governance enacted. An assessment that checks only the enterprise-wide picture will miss it, so it is worth testing the one or two workflows at the top of the list to that depth before anyone builds.

How does digital maturity connect to AI results?

The MIT research suggests the step that pays most is the one from pilots to everyday use. The 2025 update of the CISR work, covered by MIT Sloan in October 2025, found that organizations see the greatest financial impact in moving from Stage 2, building pilots and capabilities, to Stage 3, developing AI ways of working across the business. A maturity assessment is most useful when it points at that move: the workflow where a pilot can become the normal way the work gets done, with a measured result.

That move starts with a number. The workflow the assessment picks needs a baseline agreed with Finance before the build, and a baseline is a short signed document with four fields: the metric, the current number as of an agreed date, the measurement method and a verification date.

How does Future Works use a digital maturity assessment?

The workflow at the top of the assessment's list is the kind of scope a first cycle is built for: one workflow in your live operation, run for 12 weeks. In week 2 an Approved Value Baseline for that workflow is signed with your Finance team, from week 4 the work runs live, and by week 12 Finance has checked the value against the baseline. Payment for each cycle follows the outcome measured against that baseline.

Senior experts do the work, each orchestrating a fleet of AI agents on our proprietary AI-native operating system; their names go on the SOW, and a substitution needs your sign-off. The comparison with consulting firms and integrators shows where we differ.

If you already have a maturity score, start with the lowest-scoring dimension that touches a workflow with a P&L number attached, even if another dimension scores lower overall.

Keep reading

Related Articles

LLMOps is where enterprise AI quietly breaks after the demo

AI

Generative AI consulting sells slides. The P&L needs shipped software.

AI

How to Calculate ROI for a Digital Twin (Factory Model)

AI