AI maturity for knowledge-intensive organisations

Why maturity means something different for knowledge-intensive organisations

In a manufacturing company you can buy an AI solution for a concrete task and contain it. Sensors, forecasting, optimisation. The AI lives in a corner of operations and can be judged on its own output.

In a law firm, accounting practice or financial advisory there is no such corner. The core product is the judgement, documents and advice of the staff. When AI enters, it lands in the middle of the core process. It reads client documents, suggests phrasings, classifies entries, summarises case files. That means maturity cannot be measured on a single tool. It has to be measured on how AI is woven into the professional process and how well you can stand behind the output, both to the client and to a regulator.

That is the difference between having AI and being mature. The first is a purchasing decision. The second is an organisational condition.

The six dimensions

EnterpriseIQ Score measures maturity across six dimensions with different weights. The weights reflect that strategic leadership and process integration come before everything else, while culture is important but follows naturally once the other dimensions are in place.

Strategy & leadership

The heaviest weight. If executive leadership or the board do not have a clear picture of why you are using AI and what you expect to get from it, the rest falls apart. Concretely the dimension measures whether there is an articulated AI strategy with measurable KPIs, and how often the board actually discusses AI-related risks and opportunities.

Typical picture at knowledge-intensive SMEs: someone in leadership has read an article and declared "we have to do this too" without defining success criteria. Or worse: AI is delegated to the IT lead as a technical project, not a business decision. Both produce low scores because there is no ownership at the top.

Process integration

The most important signal of real maturity. How many of your core processes have AI as a measurable part of the workflow? Saying "we use ChatGPT" is not enough. The question is whether the use is deliberately integrated, measured and documented.

Example from accounting: classifying invoices automatically with AI is a process. If 80 percent of invoices are classified by AI with sample checking by the auditor, it is integrated. If someone occasionally uses ChatGPT, it is not. The difference is measurability.

Data foundation

The most often overlooked dimension and at the same time the most blocking. AI tools are only as good as the data they can reach. If contracts sit on SharePoint, client notes in Outlook threads and professional knowledge in the heads of four senior people, AI cannot draw on it.

The dimension measures whether you have a data catalogue, classification of data by sensitivity, and data owners for each core dataset. Most knowledge-intensive SMEs score low here, and it is typically the first place we recommend investing.

Technology & tools

Not "do you have many tools" but "do you have a grip on them". Number of AI tools in productive use and whether there is a formal approval process before new tools go into use. Shadow IT, where staff use AI tools on personal accounts without approval, pulls the score down.

It is common to find 10-15 AI tools in use across an organisation, of which only three are officially approved. That kind of inventory exercise is often eye-opening the first time it is done.

Governance & compliance

Do you have an AI policy that covers the use of external tools such as ChatGPT and Claude? Have you carried out an AI system inventory mapped to EU AI Act classification? This dimension grows in weight as 2 August 2026 approaches and becomes decisive after that date.

For knowledge-intensive organisations the governance dimension matters especially because you handle client data that is confidential or regulated. A low governance score is not only a compliance risk, it is a business risk when enterprise customers begin asking about your AI practice.

People & culture

How many staff have completed structured AI training? Is there an internal AI champion with time and mandate? Weights lowest of the six dimensions because culture follows naturally once the others are in place, but an entirely absent culture score points to an organisation that cannot execute on the strategy.

The typical picture is that one or two enthusiasts have learned the tools on their own time, while the rest of the firm stands still. That imbalance means the organisation cannot scale maturity into the wider team.

The four maturity levels

The EnterpriseIQ Score composite from 0 to 10 maps to four levels. Most knowledge-intensive SMEs in 2026 sit between Exploratory and Integrated. Transformative is rare and typically requires a deliberate multi-year commitment.

Moving between levels takes time. Exploratory to Established typically 3-6 months with focused effort. Established to Integrated 6-12 months. Integrated to Transformative 12-24 months. The pace is set by executive engagement and whether there is an executive AI owner with mandate and time.

Industry profiles from practice

Law firms

Typical level in 2026: Exploratory to Established. Most mid-sized law firms have staff using ChatGPT or Claude for research and first drafts, but without formal approval or policy. Confidentiality concerns often slow progress because no one has clarified which tools may receive client data.

Qualitative quick wins for the legal sector: build a shared, approved prompt library for contract review. Establish audit trail practice for AI-generated first drafts. Evaluate self-hosted models or EU residency agreements so the confidentiality question can be closed properly.

Accounting practices

Typical level: Established to Integrated. Auditors are often further along because a culture of quality control and documentation already exists in the firm. AI becomes a natural step on an existing path. Invoice classification, materiality assessments and audit notes are classic use cases that integrate relatively quickly.

Quick wins: identify the top three documentation tasks that take significant time today and could be integrated as an AI-assisted workflow. Measure baseline time before implementation. Establish sample-check procedures so the auditor always validates AI output before it leaves the firm.

Financial advisers

Typical level: Exploratory to Established, with wide variation. Credit scoring and customer segmentation are often high-risk under the EU AI Act, so maturity here is tightly linked to compliance maturity. Many financial advisers are cautious about AI because they are waiting for regulatory clarity, which means they fall behind as 2 August 2026 approaches.

Quick wins: complete an AI system inventory mapped to EU AI Act classification as the first step. Start with minimal-risk use cases (client report generation, research) where you can build experience without compliance weight. Use that experience to build governance practice before tackling high-risk areas.

IT services firms

Typical level: Established to Integrated, but with blind spots. IT services firms usually score high on the technology dimension because they use AI every day in code, but strategy and governance dimensions lag. Many build AI solutions for customers without thinking through their own practice first.

Quick wins: write a policy that covers both your own use and how you deliver AI solutions to customers. Make audit trail part of your delivery process, not an add-on. Use your own AI maturity as a sales argument with compliance-focused customers.

What you lose by waiting

"We will wait until the market settles" is a common position at knowledge-intensive SMEs in 2026. It is an understandable position, but it costs more than it looks at first.

First: the skill gap widens. The staff who use AI daily develop an intuition and speed that cannot be taught on a course. If you wait two years, competitors have a two-year lead in tacit knowledge.

Second: recruitment gets harder. Younger professionals expect their employer to have AI tools in place. A law firm or accounting practice without a formal AI practice looks dated, not conservative.

Third: when 2 August 2026 arrives, enterprise customers will start asking about your compliance maturity. If you cannot show audit trail and governance, you lose tender rounds without knowing why.

Lost momentum is hard to make up. Maturity is cumulative, not something you can buy in a quarter. The realistic risk is not doing something wrong. The realistic risk is standing still while the market moves.

Three steps you can take this week

Step 1: Measure yourselves

• Take the EnterpriseIQ Score on our homepage (12 questions, 5 minutes, free)
• Get a PDF report with your score across the six dimensions and five concrete quick wins
• Share the report with leadership and hold a 30-minute conversation about the result

Step 2: Hold a strategy session at leadership level

• 90 minutes with the executive team, no agenda beyond "where are we, where do we want to go"
• Identify 3 use cases to prioritise for the next 6 months, with a named owner
• Define 3-5 measurable KPIs that track the AI investment
• Schedule the next follow-up in the calendar, otherwise it runs out into the sand

Step 3: Write a short AI policy

• 1-2 pages, no more. Should be readable in 5 minutes
• Cover four things: a list of approved tools, rules for which data may be sent where, the approval process for new tools, and a named owner plus contact person
• Share with staff at a team meeting, not just by email
• Set a quarterly review date for the policy

These three steps build a foundation. They do not need to be perfect, they need to be in place. Once the foundation stands, you can prioritise where to go deeper next quarter, based on which dimension scores lowest for you.

FAQ

Next step

Three paths depending on where you stand: