Could the long-term winners of AI be companies that learn to use the technology more effectively than their competitors?

A follow-up to NaviOptima’s Insight: From AI Users to AI Beneficiaries – Where Is AI Already Showing Up in the Numbers?

From the Internet to AI – Who Benefits from Technological Disruption?

Over the past three decades, the internet and cloud computing have fundamentally changed how businesses operate and compete. Alongside the companies building the technology, another group of winners has emerged: businesses that have learned to use it exceptionally well.

Amazon did not invent the internet or e-commerce. Its competitive advantage came from integrating online retail, logistics, inventory management and customer data into a highly effective operating model. Amazon later became a major cloud infrastructure provider through Amazon Web Services.

Netflix, meanwhile, migrated to Amazon’s cloud infrastructure between 2008 and 2016. This involved much more than moving servers: it required a fundamental redesign of its technology architecture and software development practices. Cloud computing supported Netflix’s international expansion, although content, customer experience and its business model were also essential to its success.

New technology creates opportunities. A company’s ability to transform its operations determines how much of that opportunity becomes competitive advantage and economic value.

We may now be witnessing a similar development with artificial intelligence.

AI Adoption Does Not Yet Mean Economic Benefit

In our previous NaviOptima Insight, we examined Klarna, ServiceNow, Duolingo and Meta to explore how AI is beginning to affect productivity, growth and profitability.

We also identified a fundamental challenge for investors: financial statements report revenue, costs and cash flow, but rarely reveal how much of the change is attributable to AI.

AI adoption, operational benefits and verified financial impact are three different things.

A company may first introduce AI tools, then integrate them into everyday processes and eventually achieve measurable productivity improvements. For investors, the decisive question is whether these improvements translate into sustainable profitability and cash flow.

We now turn to three internationally operating Finnish companies: Nordea, KONE and Wärtsilä. Their potential advantage does not depend on developing the world’s best general-purpose AI models, but on combining AI with their proprietary data, expertise and business processes.

Nordea – AI as an Advantage of Scale in Banking

Banks process enormous volumes of customer interactions, payments, documents and risk-related information every day. Even a modest improvement in an individual task can become significant when repeated millions of times annually.

According to Nordea’s 2025 Capital Markets Day materials, the bank recorded 5.2 million customer conversations between October 2024 and September 2025, approximately 78% of which were handled by its Nova chatbot.

Nova uses technology from Norway-based boost.ai. According to the supplier’s own case study, Nordea operates 12 AI agents on its platform across four Nordic markets.

Nordea also uses AI in know-your-customer processes, transaction monitoring and employee productivity.

Nordea’s cost-to-income ratio, excluding regulatory fees, was 45.0% in the third quarter of 2026. The bank aims to reach 40–42% by 2030. AI is one of the enablers of this strategy, but its specific contribution to the targeted efficiency improvement has not been separately demonstrated.

NaviOptima’s Assessment: Nordea

What is the company’s difficult-to-replicate competitive advantage?

Its extensive Nordic customer base, banking data, regulatory expertise and established service processes. A chatbot can be purchased from an external supplier, but integrating it securely across a large banking organisation is considerably harder to replicate.

What financial evidence is already available?

Millions of AI-assisted customer conversations demonstrate the scale of adoption. The bank has also established a concrete long-term efficiency target. However, conversation volumes alone do not establish realised cost savings.

What evidence is still missing?

Comparable figures on working hours saved, cost per customer interaction and AI’s separately identifiable contribution to financial performance.

KONE – AI as a New Tool for Service Technicians

Elevator and escalator maintenance depends on extensive technical expertise, equipment histories and the practical experience of service technicians. When a problem occurs, identifying its cause quickly and safely is essential.

KONE’s Technician Assistant is a generative AI tool that combines equipment information, maintenance histories, technical documentation and previously resolved service cases.

In 2026, KONE reported that the tool was available to 15,000 service technicians worldwide.

It helps technicians identify possible causes of faults and potential solutions, drawing on approved technical documentation and verified historical service cases.

Rather than replacing technicians’ professional expertise, AI makes the organisation’s collective knowledge more readily accessible to them.

This is particularly relevant because much of an industrial company’s expertise is distributed across documents, information systems and the practical knowledge of experienced employees.

Potential economic benefits include faster troubleshooting, fewer unnecessary repeat visits and improved equipment availability. KONE’s 24/7 Connected Services also uses equipment data for predictive maintenance.

However, KONE has not disclosed group-level cost savings attributable specifically to Technician Assistant.

NaviOptima’s Assessment: KONE

What is the company’s difficult-to-replicate competitive advantage?

Its extensive installed equipment base, maintenance history, technical documentation and the experience of its international service organisation. A general-purpose language model does not inherently possess this specialised knowledge.

What financial evidence is already available?

Deployment to 15,000 service technicians demonstrates substantial operational reach. KONE has also described practical cases in which AI supports troubleshooting, although public information does not yet establish the average financial benefit.

What evidence is still missing?

Comparable data on service visit duration, first-time fix rates, cost savings and AI’s contribution to service business profitability.

Wärtsilä – AI for Internal Productivity and Customer Value

Wärtsilä uses AI in two economically distinct ways: improving its own operations and optimising customers’ technical systems.

According to its 2025 Annual Report, the company had deployed more than 60 AI solutions, including 25 introduced during 2025. Applications span manufacturing, research and development, maintenance, sales and administrative functions.

One concrete example is Talk–Review–Submit, developed by Wärtsilä’s own teams. It converts technicians’ spoken service reports into text and simplifies documentation.

In its 2025 Annual Report, Wärtsilä estimated the potential annual added value of Talk–Review–Submit at approximately €1 million. Wider deployment was planned for 2026. The published estimate does not in itself demonstrate realised cost savings.

Wärtsilä also uses its internally developed WärtsiläGPT and Microsoft Copilot. In customer-facing solutions, AI and advanced analytics feature in products such as Expert Insight and the GEMS Digital Energy Platform.

These solutions can help predict equipment problems, optimise system performance and reduce customers’ costs and emissions.

Wärtsilä’s most interesting long-term opportunity may arise if it can convert the ongoing economic value delivered to customers into profitable, long-term service agreements.

AI could then become not only a tool for cost efficiency but also an enabler of service business growth.

NaviOptima’s Assessment: Wärtsilä

What is the company’s difficult-to-replicate competitive advantage?

Its expertise in engines, marine technology and energy systems, combined with its installed equipment base, operational data and long-standing customer relationships. General-purpose AI cannot, on its own, replace this integrated technical expertise.

What financial evidence is already available?

Wärtsilä has published an estimate of approximately €1 million in annual added value from one internal AI solution. Its customer-facing solutions also use AI and analytics to optimise technical systems. Estimated benefits must nevertheless be distinguished from realised financial impact.

What evidence is still missing?

A breakdown of realised AI-related savings and evidence of how much value created for customers translates into Wärtsilä’s own service revenue, margins and cash flow.

European AI – Where Does Competitive Advantage Arise?

The three companies also offer insights into Europe’s technological competitiveness.

Nordea’s use of Norway-based boost.ai demonstrates that European suppliers can develop AI solutions suitable for large enterprises.

KONE’s Technician Assistant and Wärtsilä’s proprietary applications illustrate how European companies can build industry-specific AI around their own data and processes.

Wärtsilä also uses Microsoft Copilot, a US-developed product. Public information reviewed here does not provide sufficient detail about the underlying language model and cloud infrastructure suppliers used by KONE’s Technician Assistant.

It is important to distinguish between an AI application developer, the underlying language model, the cloud service and the computing infrastructure. An application developed by a European company may partly rely on US technology.

For investors, the essential question is where difficult-to-replicate expertise emerges along the value chain and who ultimately captures the economic benefits.

Europe’s AI opportunities are therefore not limited to developing large language models. They may also emerge in specialised applications and in businesses that use AI more effectively than their competitors.

What Do These Three Companies Tell Investors?

Nordea, KONE and Wärtsilä demonstrate that AI has progressed beyond experimentation into everyday business operations. However, the financial evidence varies considerably.

At Nordea, adoption is visible in millions of customer conversations. At KONE, it is reflected in a tool available to 15,000 service technicians. Wärtsilä has published a monetary estimate of the annual added value of one AI application.

These figures measure different things and are not directly comparable.

The common question is what happens next.

Will the cost of a customer interaction or maintenance task decline? Can companies serve more customers without a corresponding increase in costs? Will AI create additional customer value that supports higher prices or new revenue?

Above all, will these effects eventually become visible in operating profit and cash flow?

In our previous Insight, we outlined the AI Utilization Score concept, which examines the breadth and depth of AI adoption, process integration, automation and economic impact.

These three Finnish companies offer a useful starting point for developing that framework further. For now, however, publicly available information is insufficient for a reliable numerical comparison.

NaviOptima’s Perspective – The Next AI Winners May Emerge in Unexpected Industries

The history of the internet and cloud computing shows that technological disruption can create substantial value for both infrastructure providers and companies that use the technology.

AI may follow a similar path.

The early major beneficiaries of the AI boom have been concentrated around semiconductors, data centres, cloud services and AI models.

In the next phase, investors may increasingly focus on businesses that purchase AI capabilities and use them to transform their own operations.

These companies may not resemble traditional technology businesses. They could be banks, industrial groups, logistics providers or service companies with established market positions, valuable data and deep industry expertise.

AI can reinforce these advantages. It does not, however, automatically make a company an attractive investment. Technology costs, competitors’ progress, realised financial performance and equity valuations all matter.

From NaviOptima’s perspective, the central question remains the same as in our previous Insight:

Who builds the best business with AI?

Nordea, KONE and Wärtsilä do not yet provide a definitive answer.

They do, however, show where investors should begin looking: among companies capable of turning widely available technology into their own difficult-to-replicate and ultimately measurable economic advantage.