Using AI is no longer a competitive advantage

Almost every company now talks about artificial intelligence. More and more use it in customer service, software development, marketing, information retrieval and internal process automation. For an investor, however, that alone says very little.

The more interesting question is which companies can turn AI use into better growth, productivity and profitability. Tomorrow’s AI winners may not be the companies developing artificial intelligence. They may just as well be the companies that learn to use AI developed by others more effectively than their competitors.

AI adoption has progressed remarkably quickly. According to Stanford HAI’s 2026 AI Index, 88% of surveyed organisations used AI in at least one business function in 2025, while 70% used generative AI. Simply using AI therefore no longer differentiates companies very much. The more important question is the depth of use.

Anonymised usage data from OpenAI’s enterprise customers offers an interesting indication. Weekly ChatGPT Enterprise message volume increased roughly eightfold over a year, while the average user sent 30% more messages. Use of structured workflows grew 19-fold and reasoning-token use per organisation around 320-fold. Anthropic’s Economic Index likewise suggests that AI use is gradually moving from individual conversations towards longer-running agent-based tasks and workflows.

Companies are not merely experimenting with AI more. Some are beginning to build their operations around it.

But does it show up in the income statement?

This is where the investor’s problem begins. Financial statements tell us revenue, headcount, costs, margins and cash flow in detail. They usually do not tell us how much of the workforce uses AI every day, what proportion of workflows are AI-assisted, or how much work previously performed by people has been automated.

AI’s financial impact therefore usually cannot be isolated as a separate line in the income statement. Correlation is not causation either. If AI use increases while profitability improves, we still do not know how much of the improvement was caused by AI.

But the picture becomes more interesting when several signals appear at the same time: AI use deepens; revenue per employee rises; costs grow more slowly than revenue; margins improve; and the company can point to concrete processes in which AI has changed the way work is done.

Klarna – an unusually visible productivity shift

Klarna offers one of the most interesting examples. According to the company, revenue increased 104% from the end of 2022 to the end of 2025 while operating expenses fell 8%. Headcount declined 49% and revenue per employee rose 3.6-fold to $1.24 million. Klarna itself links the development to AI-enabled productivity improvements.

In the first quarter of 2026, revenue grew 44% year on year to $1 billion and adjusted operating income rose to $68 million from $3 million a year earlier. Revenue per employee had climbed to nearly $1.4 million.

This does not prove that the entire improvement was caused by AI. Many other factors affect Klarna. But the combination is interesting: active AI adoption + strong growth + rapidly improving employee productivity + better profitability.

ServiceNow – AI has already become business

ServiceNow is a different case. Whereas Klarna uses AI significantly to improve its own efficiency, ServiceNow sells AI as part of its customers’ workflows.

In Q2 2026, subscription revenue grew 24.5% to $3.88 billion. More importantly, ServiceNow said its AI business had exceeded $1 billion in annual contract value, while adoption of agentic AI solutions had increased ninefold in nine months.

Here the link between AI and growth is more concrete. AI is not only a tool used by ServiceNow employees; it has become a product customers pay for. At the same time, the company is targeting a roughly 31.5% non-GAAP operating margin and a 35% free-cash-flow margin for 2026.

ServiceNow therefore illustrates a second model of AI value creation: AI does not merely reduce costs — it increases the value of the product sold to the customer.

And Duolingo?

Duolingo is interesting, but somewhat different. The company uses AI to create learning content and build new interactive features. Its business model is highly scalable, so the potential cost advantage from AI-generated content and services could be significant.

In its Q2 2026 release, Duolingo reported daily active users up 23% year on year and emphasised technology as a central part of its operations. It also identifies the development and use of AI and machine learning as material to its business.

As an investor, however, I would be more cautious about drawing a direct line between AI and financial performance in Duolingo’s case. It is an interesting AI-beneficiary candidate, but not yet as clean an example of measurable economic impact as Klarna or ServiceNow. That distinction matters.

Meta shows the other side of AI investment

Meta provides another useful perspective. Q2 2026 revenue grew 28% and advertising revenue 27%. Ad impressions increased 14% and average price per ad 12%. The company itself emphasises AI as a force strengthening its core business.

At the same time, costs rose 55%, operating margin fell from 43% to 31%, and quarterly capital expenditure exceeded $31 billion.

Meta illustrates an important point: AI can improve the business while weakening short-term profitability because of massive investment. Investors should therefore not simply look for rising AI use and expanding margins. They also need to understand where the company is in its investment cycle.

Productivity deserves special attention

One simple metric deserves particular attention: revenue per employee. It is not a perfect AI measure, but together with other indicators it can be revealing.

Imagine a company whose revenue grows 15% a year while headcount grows only 3%. If margins improve at the same time and the company can demonstrate that major workflows have been automated with AI, something interesting may be happening.

A possible chain is: deeper AI use → higher employee productivity → costs grow more slowly than revenue → margins improve → cash flow strengthens.

But the chain can also run differently: AI investment rises → costs and capital needs increase first → productivity benefits emerge only later. This is why investors should follow the trend over several years rather than one quarter.

Could AI utilisation be measured?

There is no perfect metric yet, but indirect data is beginning to emerge. The Ramp AI Index, for example, tracks companies’ actual payments to AI services. In its data, the share of companies paying for AI services exceeded 50% for the first time in spring 2026.

This does not tell us how efficiently a particular listed company uses AI. But it shows that adoption can be measured through more than management commentary.

In the future, an investor could imagine an AI Utilization Score based on several dimensions: breadth of use across the organisation; depth of use from occasional assistance to daily workflows; integration into core processes; degree of automation; and the financial effect on revenue per employee, margins and cash flow.

The absolute score might be less important than the direction of change.

The next generation of AI stocks?

The first major winners of the AI boom have been builders of chips, data centres, cloud services and AI models. A second wave may follow: companies that buy increasingly powerful AI capacity and use it to transform their own businesses.

They may be banks, insurers, industrial companies, software businesses, online retailers, logistics groups or something entirely different. They may not even be called AI companies. And that is precisely why they may be interesting to investors.

Investor map – look for AI’s economic footprint

When a company talks about its AI strategy, an investor might ask: Is revenue per employee rising? Is headcount growing more slowly than revenue? Are operating and cash-flow margins improving? Is AI integrated into core operations or mainly used as a staff assistant? Is it creating new products, incremental sales or greater willingness to pay? Is there evidence of increasing AI use beyond management’s own statements? How much do the required AI investments cost? And, above all, is there already an economic footprint behind the AI narrative?

No single metric proves AI’s impact. But when several begin pointing in the same direction, the company deserves closer attention.

The NaviOptima perspective

Perhaps AI should not be viewed as a new industry, but as a new production technology.

Electricity, the internet and cloud computing eventually transformed almost every industry. The biggest long-term winners were not necessarily the original builders of the technology, but the companies that learned to use it more effectively than their competitors.

The same may happen with artificial intelligence.

So when searching for the most interesting AI investments of the future, the question may not be: Who makes the best AI?

But rather: Who builds the best business with AI?

It is a small distinction — but for an investor, an essential one.

NaviOptima | Insights

This article presents a general perspective on the theme and does not constitute investment advice. Investing always involves a risk of losing capital.