Digital Economy

AI enters the industrial phase: why 2026 will be the year of truth for CIOs

Lenovo CIO Playbook 2026 captures the turning point: 94% of European companies expect concrete returns, but the real challenge is no longer technological, it is organisational

by Marco Trabucchi

28 January 2026

4' min read

Translated by AI

Versione italiana

4' min read

Translated by AI

Versione italiana

The laboratory phase is over. Artificial intelligence in European enterprises has passed the test bench of pilot projects and is on its way to becoming a strategic business lever. But the race towards enterprise AI is now being played out on new ground: no longer who experiments first, but who manages to scale up sustainably, safely and with solid governance.

This was revealed in the Lenovo CIO Playbook 2026, conducted with IDC on 800 IT and business decision makers in Europe and the Middle East, all in companies above 1,000 employees. The numbers are clear: 94% of organisations expect positive returns from investments in AI, while almost half of the proofs of concept are already in production. Investments will grow by an average of 10 per cent over the next twelve months.

According to the most recent data, more than half of the companies have passed the test phase and introduced AI into real operations, with almost 46 per cent of proof of concept already entering production and tangible economic returns estimated at up to USD 2.79 per dollar invested in generative AI . This indicates that AI is no longer an expensive experiment, but a lever of measurable value: productivity, customer satisfaction and faster decision-making are among the most cited benefits.

The CIO Playbook 2026 paints a picture of a mature market facing structural choices. With AI becoming a pervasive component of business processes, its ability to generate value will depend on solid foundations: hybrid infrastructure, mature governance, appropriate skills, and a clear vision of AI's role in business. Shared rules, approval processes, continuous monitoring and widespread training become the real discriminating factor between who will lead the transformation and who will lag behind.

The paradox of maturity

Perhaps the most critical aspect of this new phase is governance: the set of operational rules that guarantee accountability, data security, model control and management of 'shadow AI', the unauthorised initiatives that can spontaneously emerge in organisations. And the paradox is clear: while confidence in the economic return of AI is growing, less than 30 per cent of companies have comprehensive governance frameworks in place and applied at enterprise scale. Without shared rules, AI does not scale: it increases complexity, multiplies risks and reduces the ability to control the value generated.

"Today we see clear returns from the projects invested in, but many companies still lack the skills and governance to scale AI to its full potential," explains Matt Dobrodziej, President of Europe at Lenovo. "With the move towards Agentic AI and the requirement to comply with the European AI Act, those who do not integrate trust and scalability from the start risk losing tangible returns."

Agentic AI: AI that acts

If generative AI has transformed the ability to produce content, agentic AI represents the next step: systems that do not just respond, but set goals, manage multi-step processes and can operate autonomously under human supervision. According to the research, agentic AI is the top strategic goal of CIOs for 2026, but only a minority of organisations are ready to fully scale it: 21 per cent use it today, 65 per cent say they are ready to scale it in their operations within twelve months, but adoption remains at an exploratory stage for most. The challenge here is twofold: technological and control. Automating complex processes raises questions about human oversight, accountability and risk management that require clear rules. But the way is now clear. According to some IDC forecasts, agentic AI will not remain confined to prototypes: growth is expected to double the productivity of the workforce by 2027, exceeding the current impact of generative AI.

Hybrid as standard

On the infrastructure front, a model is being consolidated: the hybrid architecture. 82 per cent of organisations now see AI in a hybrid context combining public cloud, on-premise infrastructure and edge computing, balancing performance requirements, cost, data sovereignty and compliance. In practice, AI needs to be where the data is and where value is needed, not just in the cloud. At the recent CES 2026, Lenovo expanded on this vision by introducing the Lenovo Hybrid AI Advantage platform and new solutions that simplify AI execution in hybrid environments, with new inference tools, intelligent agents and automation of AI operations across servers, devices and the cloud.

Italy among the leaders

Italy ranks among the top European countries. 74% of Italian companies are at an advanced stage of adoption, above the regional average. 90% plan to increase investments in the coming year. The first tangible returns are being recorded in IT, data analytics, cybersecurity and marketing. The focus on security, automation and internal skills development is growing, with almost a quarter of Italian companies ready to focus on Agentic AI. It should be noted, however, that the snapshot only concerns the enterprise world: all the organisations involved have more than 1,000 employees, with a significant proportion above 10,000. SMEs, which represent the backbone of the Italian production system, operate with different dynamics, resources and structures.