The commercial success of artificial intelligence developers and providers does not by itself ensure that companies using the technology create value, Marco Mongiello said as the first speaker at the HVG AI Talks conference. In his view, organizations must do more than introduce new tools: they must reshape how they operate and redesign workflows around clear goals.
He identified trust in the technology and the skills needed to use it as two conditions for creating value. As an example, he described a manager who spends a few seconds instructing AI to produce a report, only for an employee to spend hours checking it. In that case, the work may not become more efficient; it may simply be shifted to someone else. Employees therefore need to understand the tools’ capabilities and limitations.
Three types of workflow, different risks
Mongiello said companies should divide their workflows into three groups. Some tasks can be accelerated with AI at relatively low risk, others need to be fundamentally redesigned, and there are areas where the technology should not be used.
As an example of the risks, he cited a report prepared by Deloitte for the Australian government at a cost of several hundred thousand dollars. Following the use of AI, the document contained nonexistent scientific references and other serious errors. An Air Canada chatbot gave a passenger incorrect information; the airline tried to disclaim responsibility, but a Canadian dispute-resolution body ruled in the passenger’s favor.
As a positive example, Mongiello pointed to a partnership between OTP Bank and Mastercard, under which a live card transaction initiated by an AI agent has already been completed in Hungary. He said AI’s role is not to replace people but to multiply their capabilities. To achieve that, companies must balance efficiency gains against safety.



