The Power of sharing
Subscribe to ViewPoint and get curated content delivered to your inbox.
Sign up© Opticos 2025. All rights reserved.
Artificial intelligence (AI) remains one of the most significant opportunities for large organisations. Yet, despite record investment and widespread enthusiasm, most companies remain at an early stage of maturity. Across Europe, over a third of enterprises report some form of AI adoption, but only a small fraction has successfully scaled AI to deliver tangible business value(1). Recent studies show that nearly three-quarters of companies struggle to move beyond isolated pilots, with as many as 85% of AI initiatives falling short of their objectives(2). This gap between intent and impact is the defining challenge of the current AI era.
This article draws on Opticos’ experience working with leading organisations in Sweden and across Europe, as well as industry practices, to analyse the root causes of AI project failure and outline practical steps for achieving sustained results at scale.
The proliferation of AI pilots reflects growing recognition of AI’s potential to drive productivity, growth, and innovation. However, most companies still find themselves “stuck in pilot mode.” Few have succeeded in integrating AI into core business processes or realised measurable improvements in efficiency, customer experience, or profitability.
This pattern is consistent regardless of sector. The enthusiasm for AI as a transformative force has led to substantial experimentation, but the transition from proof-of-concept to enterprise-wide impact remains elusive for the majority. Data shows that just 1%(3) of companies describe themselves as fully AI-mature, with AI embedded into workflows and driving significant, sustained value.
Analysis of recent failures highlights several recurring obstacles. Addressing these is crucial for moving beyond experimentation toward real operational and strategic impact.
Many AI projects start as tech-driven experiments with only minimal business input. When business and IT work in separate silos, they struggle to agree on the right problems, metrics, and sources of value. Data-science teams may produce brilliant models, yet these often miss the mark because they were built without day-to-day operational insight. Without cross-functional collaboration, initiatives rarely develop a clear hypothesis tied to outcomes such as revenue growth, churn reduction, or cost efficiency. Expectations drift apart, objectives grow fuzzy and returns disappoint, sometimes leaving the business with a set of promising pilots that cannot scale because IT and operations are out of sync. Executive sponsorship is hard to secure until both sides back a shared value case, especially when the benefit, like saving employee time, is tough to measure in hard currency.
A counterexample comes from H&M. There, merchandisers, supply-chain planners, and data specialists joined forces to create an AI demand-forecasting platform. By pooling expertise from the start, the team cut excess inventory by roughly 15 percent and increased sales by about 10 percent, showing how close business-IT cooperation can turn data science into real sourcing gains(4).
AI systems are only as good as the data that powers them. Poor data quality, absence of relevant data, fragmented sources, and lack of integration are among the most common reasons AI projects fail to scale. Incomplete, outdated, or inconsistent data undermines model performance and erodes trust among business users. In your AI project, this data question can be the most tricky and costly one. Data can put serious limitations on your AI ambition level.
As AI adoption grows, so too does the importance of strong governance frameworks. In Europe, regulatory requirements such as the GDPR and AI Act demand rigorous oversight of how data is used, how models make decisions, and how ethical risks are managed. Many companies are unprepared for this environment, lacking formal structures for accountability, transparency, and compliance. This leads to delays, unanticipated risks, and in some cases, the halting of promising pilots due to privacy, fairness, or security concerns. Governance is no longer optional, it is a prerequisite for sustained AI success.
AI transformations are as much about people as they are about technology, if not more. While technical capability is essential, the greatest challenges often arise on the organisational and cultural side. Employees may lack trust in new AI tools or fear displacement; middle managers may resist changes to established processes. Crucially, leadership inertia remains a significant barrier. Without visible executive sponsorship, clear communication, and a commitment to reskilling, AI projects frequently stall at the pilot stage. Success depends on leadership driving the transformation agenda and modelling the behaviors required for adoption.
Leading companies recognize that effective AI scaling requires investment beyond algorithms and infrastructure. A common pitfall is over-investing in model development while under-investing in the change management, data readiness, and business transformation required to onboard AI into day-to-day operations. Evidence suggests that approximately 70% of the effort in successful AI transformations is devoted to people and processes, including training, business redesign, and workflow integration, while only 10% is focused on the underlying algorithms(5). Those that achieve the greatest value from AI allocate resources accordingly, ensuring the environment supports sustained use and improvement.
Drawing on our experience and analysis of the most advanced adopters, Opticos has identified four practices that distinguish organisations able to scale AI for real business impact:
Effective AI initiatives start with a clear definition of the problem or opportunity. Leaders identify where AI can move the needle on key performance indicators, whether increasing revenue, improving customer satisfaction, or reducing operational costs. Each initiative is justified by a measurable value hypothesis, and early efforts focus on high-impact, feasible use cases to build momentum. This ensures that technology investment remains closely tethered to business outcomes, and success is defined in terms the business understands.
AI work requires close collaboration between business experts and technical teams. Top companies form stable, cross-functional squads made up of domain specialists, data scientists, engineers and IT staff, and keep them together from the planning stage all the way through deployment. A central role in these squads is the analytics translator. This person turns business goals into clear technical requirements and makes technical results easy for everyone to understand and act on. This way of working speeds up delivery, raises adoption rates and cuts the chance of costly mistakes. It also calls for a smart sourcing strategy. Bringing in outside partners for specific tools, platforms or skills helps the core team stay focused on value creation and keeps total costs clear.
Success at scale depends on trustworthy, accessible data and a governance framework that supports both innovation and compliance. Leaders invest in data infrastructure, consolidation, and quality management early in the journey, often establishing a central data office or council to set policies and resolve issues. Governance frameworks clearly define ownership, access rights, and standards for responsible AI development, including privacy, fairness, and transparency requirements. In the EU, alignment with GDPR and evolving AI regulations is a must. Forward-looking organizations treat compliance as an enabler of trust, not a barrier to progress.
AI adoption is accelerated when supported by a culture that values experimentation, continuous learning, and transparency. Senior leaders communicate a compelling vision for AI, set expectations about the opportunities and risks, and champion ongoing education for all employees. Change management efforts focus on involving end-users early, offering role-specific training, and demonstrating early wins to overcome scepticism. Many leading firms invest heavily in upskilling, not only technical talent, but also business leaders and frontline employees, ensuring that everyone can engage effectively with AI solutions and identify new opportunities for value creation.
Recent Nordic case studies illustrate these principles in action. For example, a leading fintech company has automated the majority of its customer service inquiries using AI assistants, resulting in significantly faster resolution times and millions of euros in annual savings. The keys to success were clear business-driven use cases, close collaboration between product and AI teams, rigorous data management, and a robust change management program that engaged customer service representatives in the design and rollout of the solution(6).
In manufacturing, several Swedish firms have successfully implemented AI-driven predictive maintenance, reducing equipment downtime and optimising maintenance costs. These organisations invested early in high-quality, accessible data and ensured that technicians were involved in the design and interpretation of model outputs, resulting in higher adoption and clear, measurable impact(7, 8, 9).
AI has evolved from a speculative experiment into a strategic engine for business change. As regulators tighten rules and stakeholders demand transparency, scaling AI safely and responsibly is now a key competitive edge. Success depends on closing the gap between strategy and technology, investing in data and people, and embedding governance, ethics and sustainability throughout your programs.
To turn AI promise into real value, define outcomes in business terms first, then stand up cross-disciplinary teams of domain experts, data specialists, engineers and translators. Build strong data foundations and oversight, plan for change through training and clear communication, and bake in ethical and transparency standards at every step. This unified approach transforms pilots into a lasting advantage.
At Opticos, we support leadership teams across Sweden and Europe in navigating this journey, applying a disciplined, use-case-driven approach, fostering business-IT alignment, and ensuring that AI investments deliver real, sustainable outcomes.
Contact Opticos to explore how we can support your transformation journey.
Contact Us
Tatiana Schön, Ankita Bhargava, Rickard Holmkvist & Johan Valeur
