Ai costs millions and there is no effects. You can feel disappointed in business

About artificial intelligence, which was to revolutionize business and brought disappointment. About why companies do not see the effects of AI and how to change it in an interview with Radosław Mechło from Buzzcenter.

Beata Anna Święcicka, “Wprost”: You have dozens of successful AI implementation in companies and several thousand trained people. Such a lot of experience sharpens the look, so it is impossible to start our conversation with the basic question: how does business deal with AI?

Radosław Mechło, Head of Ai in Buzzcene: I will answer, recalling the May article from “The Economist” “Welcome to the ai trough of disillmentment” (“Welcome to the AI ​​disappointment valley” – translated by ed.). It cited a bitter confession that the president of a technology company from San Francisco heard from the head of one of his clients from the Fortune 500 list: “I don’t understand why it lasts so much. I spent money on it, and it doesn’t work.”

For many enterprises, the initial excitement with the potential of generative artificial intelligence actually gave way to disappointment associated with the difficulties in translating this technology into real productivity.

According to S&P Global Market Intelligence, cited in the said article, it appears that the percentage of companies abandoning most of its pilot Genai projects increased this year to 42 percent. Only a year earlier it was 17 percent. In other words, almost half of the companies accused most of the AI ​​initiatives, without seeing the results. Even bold projects are inclined to review the approach: President of FinTech Klarana recently admitted that he has gone too far, replacing people with the AI ​​system in customer service and now he must employ staff again. In Europe, as the AWS report “The State of Ai in Europe” (2025), only 14 percent companies recognize that AI generates real business value, although as much as 42 percent. declares her regular use.

So business uses, but does not gain. Why? What are the main reasons that stop companies from being transformed by AI’s potential into real business benefits?

Statistics signal that after a wave of enthusiasm towards AI, many boards are experiencing today what Gartner calls the “valley of disappointment.” It is a phenomenon that appears regularly in the IT world, when after Hype and the accompanying euphoria, the inflated expectations collide with reality.

The desire to jump at AI was huge, the investments flowed in a wide stream, but in numerous cases the results turned out to be meager.

This is well illustrated by the results of the Boston Consulting Group analysis from last year. The company determined that only 26 percent enterprises have developed competences to go beyond the pilot stage and generate a measurable value thanks to AI. The others, and therefore nearly three -quarters, still did not achieve tangible effects. Similar conclusions come from the already quoted S&P Global report, which shows that the statistical organization abandoned 46 percent. AI’s initiatives are already at the Proof-V-Concept stage before they went into production. In other words, almost every second promising prototype has never been implemented.

What is behind this wave of disappointment? Is business not ready for AI?

This is a combination of several reasons: financial, technological, organizational and regulatory. Costs and no expected return on investment come out on the forehead. The training of AI generative models and maintaining adequate infrastructure can be horrendously expensive – we are talking about millions of outlays – so when the promised benefits do not materialize, CFO frustration and budgets’ cutting are understandable. In the technological survey of CDW, almost two -thirds of respondents assessed that they received a refund of at most 50 percent from AI. investments incurred, and the full return (100 %) reached less than 2 percent. companies.

Almost nobody earns on AI yet, and most add. No wonder that enthusiasm is weakening. Still, a certain dose of optimism remains.

According to the IBM report from 2024, most companies that have not yet reached a full ROI with AI provide for significant savings over the next three years, and almost half expect clear financial benefits by 2027.

So what about Roi? Is there a chance for Roi with AI?

Lack of roi often results from unclear business purposes. Many companies have “rushed” on artificial intelligence under the pressure of fashion, without a thoughtful plan, how this technology is to specifically increase revenues or reduce costs. Fortune magazine said in October 2024 that nearly 75 percent Corporate AI initiatives did not fulfill their promises precisely because they were implemented quickly, without proper connection with business processes. The technology was focused instead of the problem to solve, and then it was surprised that the return on the investment is foggy.

The difficulties in integration of AI solutions with existing processes and applications are also significant. Often, what worked great in the R&D section cannot “get along” with the production system or company workflow. Integration can be expensive and complex, and the lack of compatibility discourages teams from continuing. In addition, there is a purely infrastructure issue: computing power restrictions.

The demand for power to train models exceeded the supply. Over 80 percent Companies, according to the CIVO report, indicated the deficiencies of GPU processors as the main reason for delays in the implementation of AI. Many initiatives got stuck in the testing phase, waiting for infrastructure and available chips to keep up with ambitions. In the background we also have regulatory risks and security issues.

And the regulation issues mentioned by you? How do they affect the implementation of AI?

Uncertainty about the shape of future AI regulations, especially in Europe, certainly cools enthusiasm to wider implementation. The AWS report “The State of Ai in Europe 2025” shows that companies operating in conditions of increased regulatory uncertainty invest an average of 28 percent. less in artificial intelligence. There is a fear of slip -up: many CEOs are afraid that the immature AI system will make a mistake costing the company’s reputation or exposes it to the allegation of violation of privacy. The more regulated the industry (e.g. financial or health), the more careful you approach the experiments with AI. So many organizations suspend projects until the legal environment and the standards of dealing with AI become more clear.

Hence the disappointment AI?

The cumulative effect of the above barriers is present tiredness and skepticism towards AI in business. This phenomenon is not unusual, however. According to the concept of the Gartner hype cycle, after the initial “peak of inflated expectations” there is a disappointment phase when reality does not match Hype.

Generative AI caused euphoria in 2023 – each company wanted to have its own chatbot or “Copilot”, often acting under Fomo pressure. A year later, excitement gives way to difficult questions: what went wrong and where did the promised effects go?

In the article “Welcome to the Ai Trough of Disillusionment” analyst Gartner John Lovelock notes that many managers are just sliding into this valley of disappointment. Gartner predicts that it will last until around 2026 – that is, for the next dozen or so months business will digest the experience of unsuccessful AI initiatives and look for better roads. At the same time, the giants, what Alphabet, Amazon, Microsoft, the finish line does not stop investing huge amounts in the development of algorithms and infrastructure. As Bain & Company indicates, the outlays on AI are record -breaking today, but skepticism about their return is growing. Technological giants, however, not only preach AI’s potential, but also use it at home: Google, Microsoft, Meta or Amazon saturate their own products and processes with intelligent solutions, counting on the performance jump and new sources of advantage. If these efforts are successful, they can again awaken the faith in other companies that AI, however, can bring real benefits.

So how companies outside the BIG Tech industry should react to this cold shower. Is it offended by AI or can he draw conclusions and prepare the ground for another wave of innovation?

The experience of market leaders suggest that the path to success leads by improving the organization, not escaping from technology. The BAIN consulting company emphasizes that the very implementation of AI does not bring great benefits.

Real results only appear when the company changes the way of working, the structure of action and develops the competence of the team. To put it simply: without people who can use AI and without adapting their work to its capabilities, even the best algorithm will not change much.

Research and practice clearly show that competence cannot be omitted. The Asana report from 2025 revealed that more than half of the AII questioned, regrets the implementation of AI without proper training of employees, and a third even withdraws or lists the implemented AI solutions for other, more suited to the realities of the organization. Such a reflection means one thing: Hype has fallen, it is time for hard work at the root. Companies that now invest in education and development of staff, ordering data and invest in infrastructure, have a chance to break free from the “valley of disappointment” and enter the “slope of enlightenment”.

Patience, therefore, therefore remains. Is it payable to companies?

Paradoxically, the current cooling of moods can come to business for good, provided that it is properly used.

When the fashion for superficial implementation for the very possession of AI is over, enterprises will gain significance that will build solid competences and approach artificial intelligence strategically. Ai will mature – like any breakthrough technology – and those who do not lose patience and prepare their organizations wisely, collect fruit.

Gartner predicts that after the current valley the stage of real productivity will come, when mature applications of AI bring measurable benefits. Then those skeptics who are hastily discouraged today may regret the omission.

To sum up, what should companies do today to prepare for the future with AI?

There is one conclusion: education and organizational competences are the key to the transition from hype to real business effects. Management boards of companies that will now focus on the development of internal competences – from training with AI, to the construction of the right infrastructure and culture based on analytics – significantly increase the chances that artificial intelligence will fulfill their promises: maybe not immediately with fireworks, but gradually, in a permanent way and felt in the company’s results.

Thank you for the interview.

Bio: Radosław Mechło – consultant, trainer and trainer who effectively combines technological competence with business experience. Cybernetic by profession, with over 15 years of management experience. Co -founder of Buzzcenter – a company specializing in consulting and education in the field of artificial intelligence. He collaborated with market leaders, implementing advisory and training projects for such brands as: Danone, Polpharma, Bayer, Ipsen, Classen, Kotányi, Streamsoft, Exatel, Ikano Bank, Vastint, Warta, Bank BPS, Leroy Merlin, Zaiks, Budimex, TPa. He specializes in digital transformation of enterprises, helping managers and teams to maximize the benefits of AI implementation.