What other companies and experts think
Posted: Wed Jan 22, 2025 8:53 am
vide. "To solve this problem, we thoroughly justify the potential economic effect and agree on it with all participants in the process. Then, together with the customer, we pilot the solution, confirm the effect, and only then does full-scale implementation and large investments begin," explained Alexey Bogomolov. He added that, in kenya whatsapp number database addition to this, Reksoft faces the problem of data quality and availability. In this case, the company works with the client's employees to structure and describe the data or offers projects to modernize the IT landscape in terms of data and analytics, and build a data management system.
Mikhail Tregubenko, Director of Organizational Projects at PAO SIBUR Holding (an oil and gas chemical company), agreed that building new development processes, developing talent, and improving data quality are important focuses when implementing AI. He also identified two more challenges: protecting confidential information and vulnerabilities in systems, as well as limiting existing systems to integration with AI.
"We are actively working on the potential for implementing AI at SIBUR. Perhaps the key task is to bring the current level of digital maturity to the level required for implementing AI in most processes," said Mikhail Tregubenko.
The experts interviewed by ComNews agree with the findings of the study and confirm that the problems mentioned exist and are important. Georgy Banchikov, CEO of the developer Ingry.Tech LLC, believes that it is difficult to integrate artificial intelligence into existing business processes and that it is necessary to treat the implementation of AI as research and development work, and be prepared for negative results. "Such work is constant reengineering and changes. At the same time, it is not a fact that each change will be successful, so fatigue from such innovative activities accumulates very quickly. It does not matter whether the company is large or small - the integration of AI into work will still lead to additional costs," says Georgy Banchikov.
He added ethical issues to the problems, such as data security and information security: to obtain results, data must be downloaded, so you may encounter issues of information security, as well as validation of results. "It's one thing when someone writes code and puts a certain logic into it: then it's clear who is responsible for it. If artificial intelligence does this, then it's unclear whose responsibility it is," notes Georgy Banchikov.
Mikhail Tregubenko, Director of Organizational Projects at PAO SIBUR Holding (an oil and gas chemical company), agreed that building new development processes, developing talent, and improving data quality are important focuses when implementing AI. He also identified two more challenges: protecting confidential information and vulnerabilities in systems, as well as limiting existing systems to integration with AI.
"We are actively working on the potential for implementing AI at SIBUR. Perhaps the key task is to bring the current level of digital maturity to the level required for implementing AI in most processes," said Mikhail Tregubenko.
The experts interviewed by ComNews agree with the findings of the study and confirm that the problems mentioned exist and are important. Georgy Banchikov, CEO of the developer Ingry.Tech LLC, believes that it is difficult to integrate artificial intelligence into existing business processes and that it is necessary to treat the implementation of AI as research and development work, and be prepared for negative results. "Such work is constant reengineering and changes. At the same time, it is not a fact that each change will be successful, so fatigue from such innovative activities accumulates very quickly. It does not matter whether the company is large or small - the integration of AI into work will still lead to additional costs," says Georgy Banchikov.
He added ethical issues to the problems, such as data security and information security: to obtain results, data must be downloaded, so you may encounter issues of information security, as well as validation of results. "It's one thing when someone writes code and puts a certain logic into it: then it's clear who is responsible for it. If artificial intelligence does this, then it's unclear whose responsibility it is," notes Georgy Banchikov.