AI: useful and safe

“Isn’t artificial intelligence taking over new areas too quickly? Will it leave people without jobs? – such questions concern everyone who follows the development of modern technologies. How is AI being implemented in Tajikistan? Let's talk about this using the example of one of the technology leaders in the telecom industry. Spoiler: even where AI works, the last word still remains with humans.

A look at the day after tomorrow

The need to control the development of AI, including in our country, is obvious. It is no coincidence that Tajikistan took the initiative to establish an International Day of Safe, Secure, and Trustworthy Artificial Intelligence - by a UN decision, it will be observed annually on August 31. This is a strategic look toward the future—a concern not so much for today, but for the day after tomorrow. Even now, however, Tajik companies are successfully implementing AI technologies. But how high is the risk that they might completely replace humans?

According to experts at MegaFon Tajikistan, this danger is greatly exaggerated. AI acts as an assistant that performs routine tasks more quickly, thereby increasing the efficiency of the people who assign tasks to it. Concrete examples of AI implementation at the company include the "Dono" chatbot, which handles customer service, and the development of the multifunctional mobile app "MegaFon Life."

Ask Dono for the answer

Every day, subscribers ask their mobile operator tens of thousands of questions. Many of these, however, are routine. For instance, customers frequently ask to check their current account balance, inquire about how many gigabytes remain in their data package, or seek advice on which new tariff plans might be the most advantageous for them. An AI assistant is well-suited to provide this kind of information—which is exactly how Dono was developed. It can recognize requests made in natural, conversational language, allows users to choose between Russian and Tajik, and can resolve many issues without human intervention.

Dono is highly efficient. For instance, when subscribers reach out regarding the activation of eSIMs, the intelligent assistant handles over 60% of these inquiries entirely on its own: it provides the necessary information, explains the activation process, and lists the documents required to set up an eSIM. However, the actual processing of the eSIM setup is carried out by human staff.

Across all topics, Dono handles more than 50% of inquiries—meaning that over half of the requests reaching the online service channel are resolved directly by the chatbot. This saves time for human consultants, reduces the volume of routine queries, and provides a more convenient path for subscribers to access services.

Dono is constantly learning and recognizing an ever-growing number of requests, having become a full-fledged tool for the digital transformation of customer service. And yet, subscribers always retain the option to connect with a human consultant, who can resolve even the most complex issue. For some residents of Tajikistan, interacting exclusively with a human being remains a fundamental preference. Consequently, AI will not entirely replace people in the service sector. However, it is an undeniable fact that it serves as a highly valuable and effective assistant.

AI helps IT

The same can be said for the IT industry. The idea of ​​AI independently conceiving, developing, and implementing a product is more the stuff of science fiction. Neural networks rely on existing data, and they are not always able to distinguish between reliable information and fakes. This is precisely why human traits—such as critical thinking and a creative approach—are so important.

For instance, when it comes to developing the "MegaFon Life" app, it is people who come up with new features and decide what will be necessary and useful for subscribers. Artificial intelligence is used as a tool—for processing information and carrying out routine operations and standard tasks. This allows specialists to focus on more complex issues that require expert knowledge, and to devote more attention to analysis and decision-making. As a result, the team can move more quickly from defining a task to its implementation, and can spend more time on product development and exploring new solutions.

AI safety is ensured through mandatory oversight by responsible developers. Team specialists monitor the solutions and code generated by the AI, conducting regular code reviews—checks for the quality and accuracy of the software code—before any changes are implemented in the application. This is how innovative digital services for MegaFon Tajikistan subscribers are created: convenient, intuitive, and secure.