The obstacle is not simply access to technology
Eurostat reports that 20% of EU enterprises with at least 10 people employed used AI in 2025. Adoption was around 55% among large enterprises, compared with roughly 17% among small businesses. The difference is striking, but adoption alone tells us little about business value.
The more revealing figures concern businesses that considered using AI but did not proceed. Within that specific group, 70.3% cited a lack of relevant expertise, 53.6% uncertainty about legal consequences and 52.7% concerns about data protection and privacy. Only 17.8% believed AI would not be useful to their business. These percentages do not describe all EU companies.
For many leadership teams, the missing ingredient is not another demonstration of what AI can do. It is a clear way to decide which opportunities are worth pursuing, how to manage risk and how success will be measured.
AI needs management ownership and board oversight
AI decisions affect costs, service quality, how work is organised, operational risk and customer relationships. They cannot be delegated entirely to IT. Management needs to set priorities and appoint business owners; boards need to scrutinise value, controls and exposure.
A 2025 Deloitte survey found that 66% of respondents regarded their boards’ AI knowledge or experience as limited or non-existent, while 31% said AI was not on the board agenda. These are respondents’ assessments, rather than an objective test of board capability, but the governance gap deserves attention.
What can we learn from Polish organisations?
A 2026 report by Anna Majkowska and Joanna Tokar, “Future Skills and AI: Organisational Readiness for the Future of Work”, drew on responses from representatives of 134 organisations. While 94% of respondents viewed AI as an opportunity, only 16% of the organisations surveyed treated it as part of their broader development direction.
The gap between enthusiasm and structured action matters. The report examines leadership, capabilities, processes, data and adaptability, not merely whether a company has bought an AI tool. Its sample was not designed to represent all Polish mid-sized businesses, so the findings should be read as insight into the participating organisations, not as national prevalence estimates.
Why start with management, not infrastructure?
A mid-sized company does not necessarily need a new platform, a sizeable engineering team or a major systems overhaul as its first move. It often needs a disciplined management conversation: which processes matter, what they cost today and who will own the improvement.
In my experience leading transformation programmes, the volume of ideas is far less important than the organisation’s ability to select and deliver the few that matter. A strong programme makes it possible to reject weak propositions early and focus resources on applications with a credible business case. This draws on the author’s professional experience; it is not presented as a verified result of a specific Symmetria Partners engagement.
Four things managers need to understand
- First, where AI may create measurable business value. Start with process performance: turnaround time, error rates, service quality, cost, revenue and customer experience — not the latest model.
- Second, what employees are permitted to do. Establish approved tools, data-handling rules, human review and decision rights. Under Article 4 of the EU AI Act, as amended by Regulation (EU) 2026/1744, AI providers and deployers must take measures to support AI literacy among staff and others operating or using AI systems on their behalf. A short training programme and a practical policy can help, but they do not replace an assessment of the legal requirements applicable to each use case.
- Third, how to run a pilot that ends in a decision. Every pilot needs a business owner, a measured starting point, an expected benefit, a review date and clear criteria to scale, revise or stop.
- Fourth, how to bring people into the change. Employees need to understand the purpose of the application, the rules for using it and what will change in their work. Otherwise, even technically sound tools may see little sustained adoption.
A practical starting point for mid-sized firms
Use one leadership session to identify five processes that matter most for customer outcomes, costs or growth. For each, document the problem, today’s baseline and the potential role of AI.
Then review which AI tools people already use, including what information they share with them. Set proportionate rules and make it clear who approves new tools and use cases.
Choose one application whose business impact can be measured. Give it an owner, a budget and a decision date. Review the evidence before committing to broader deployment.
The management decision comes first
Mid-sized companies do not have to begin with a sweeping technology programme. They can begin with a business decision: why use AI, who is accountable, how will data be protected and what evidence will demonstrate value? Management capabilities do not replace technology. They make it useful.
Assess your organisation’s AI readiness or speak to us about where to begin.
Sources and further reading
- Eurostat, AI use by enterprises (2025)
- Eurostat, The use of AI technologies in the EU (2026 analytical report)
- Deloitte, Governance of AI: A Critical Imperative for Today’s Boards (2025)
- Majkowska & Tokar, Kompetencje przyszłości i AI (2026)
- EU Artificial Intelligence Act, consolidated text
- Regulation (EU) 2026/1744
