The artificial intelligence market is currently going through a kind of truth detector. We are leaving behind the time when any AI initiative was financed in advance, on the general credit of trust in the technology as such. Now every investment must be backed by concrete figures in a financial report, not by a persuasive presentation.

In short, here is what I see on the market right now:

  • ROI without a comparison to the cost of capital is a meaningless number;
  • Most failed AI projects failed not because of the technology, but because of the approach to implementing it;
  • Budgets are shifting from the periphery to the core of the business, and this is temporarily putting pressure on the market's average ROI;
  • The right metric of success is the financial result, not the quality of the presentation;
  • The next stage is the transition to agentic AI as an engine of growth, not just an assistant.

A 7% ROI as a loss of shareholders' money

A 7% return sounds decent on its own. But business does not exist in a vacuum. If a company's cost of capital is around 10% and a project brings in 7%, you are effectively burning shareholders' money, even if there is some kind of plus on paper.

ROI (Return on Investment) – is a measure of the profitability of investments, that is, the ratio of the profit (or benefit) from an investment to the amount invested.

I see many executives continuing to report a positive ROI without mentioning this gap. That is not a financial report, it is a marketing document, – Artem Lyashanov.

Why 75% of initiatives fail to meet expectations

Most companies whose AI projects failed to meet profitability expectations fell into the same trap. They bought the technology as a fashion accessory rather than as a financial instrument. When a project stays on the periphery of the business, outside key operational processes, it physically cannot deliver a large-scale effect, no matter how impressive the algorithm itself may be.

To executives who are under pressure today because of the low returns of their AI initiatives, I give one piece of advice. Stop reporting on successful tests and start reporting on financial results, – Artem Lyashanov.

Why a systemic approach temporarily lowers ROI

We are seeing a large-scale shift in budgets. From peripheral experiments to the core of companies' operational processes. Working with the periphery was easy, it delivered a quick but limited result. Moving to the core of the system is a much more complex task. It requires deep coordination between functions and work with the company's fundamental data.

It is precisely this complexity that explains the temporary decline in the market's average ROI. But it is here, and not on the periphery, that the potential for real scaling lies, rather than for cosmetic fixes.

What comes next

The next step in this evolution is the transition to agentic AI. We are building the foundation for an autonomous business, where technology stops being just an assistant and becomes one of the main engines of growth.

Agentic AI – is the next stage in the development of artificial intelligence after chatbots and AI assistants.

The AI gold rush has turned into a game of survival of strategies. The drop in ROI to 7% is a natural consequence of moving out of comfortable, peripheral positions into the complex core of the business. Today, leadership is measured not by the number of neural networks on staff, but by the ability to turn the autonomy of agentic AI into real profit for shareholders.

This material is for informational purposes only and reflects an expert assessment of industry trends as of 2026.