Advisors do not always take root at enterprises
Posted: Wed Jan 22, 2025 3:50 am
Evgeny Burkhan said that an autonomous plant can increase asset productivity by 1-3%, reduce losses during transient processes by 10%, and reduce variable costs by 6%.
Sergey Mukhantsev, Director of Product Development at AiLine PJSC Softline (Softline Digital), reported that, based on the company's experience, the average lifespan of advisors in production is three to four months - after which the operator does not contact them.
"Advisors give increases azerbaijan whatsapp number database of around 1-2% - that's already good - and a separate question is how to fix them. Another problem is when you tell a technologist that you will give a 5% increase, he answers: "They will raise my plan by 5%. This thing will break for you, what will I do next year? Let's at least agree on 2%. " That is, there is a human factor," Sergey Mukhantsev shared his experience.
for other reasons. Andrey Golov, Head of the Machine Learning and Artificial Intelligence Center at Severstal Digital, explained why this may happen. "There are people who say that they have worked for 20-40 years and know what to do better than an advisor. Sometimes this turns out to be true. An advisor needs to be modernized and improved continuously," Andrey Golov explained.
There are also cases when an employee who is busy with a lot of work does not have time to set up the advisor's work mode, so the company does not receive the effect of the models.
The third reason is that employees are not ready to work with the technology - they want to see a high-quality autopilot and will not settle for anything less. "If they see a simple cruise control, they refuse to use it. Therefore, the model has no effect if we do not switch to an autopilot," explained Andrey Golov, adding that, nevertheless, the use of advisors is justified in some cases.
Some participants in the discussion said that it would be better to move on to autopilots right away. Andrey Golov said that at their enterprise, most decisions are models that make control actions with a short periodicity - a second, a minute. "All models are actively involved in the process. Advisors are dying out, we have few of them left. We are aiming only at autopilots. If it is not an autopilot, then employees stop using the decisions," Andrey Golov said.
However, the participants noted the problem of the high cost of autopilots, which are expensive to implement. "Advisors are a very good start for further development for the SUUTP (improved process control system - ComNews note )," said Danila Malina, head of production digitalization at Uralchem. "It is expensive to immediately install an autopilot."
The participants in the discussion agreed that one does not exclude the other and that both autopilot and digital advisors can be used at the enterprise.
"It is more effective to use both approaches. Where autopilot is possible, use it. Where it is difficult to justify the costs of implementing additional automation, where it is not worth it or there is an operator, use this option," said Emil Gasanov, software product development manager at Nauka LLC.
At the plenary session, Deputy Director for Digitalization of JSC MCC EuroChem Valery Cherepanov said that, despite the movement towards automation, people will be needed at enterprises. "People will focus on other tasks. Therefore, we mean that unmanned production is possible, but we are unlikely to get there. We need those who will develop, maintain, and develop these systems," said Valery Cherepanov.
Sergey Mukhantsev, Director of Product Development at AiLine PJSC Softline (Softline Digital), reported that, based on the company's experience, the average lifespan of advisors in production is three to four months - after which the operator does not contact them.
"Advisors give increases azerbaijan whatsapp number database of around 1-2% - that's already good - and a separate question is how to fix them. Another problem is when you tell a technologist that you will give a 5% increase, he answers: "They will raise my plan by 5%. This thing will break for you, what will I do next year? Let's at least agree on 2%. " That is, there is a human factor," Sergey Mukhantsev shared his experience.
for other reasons. Andrey Golov, Head of the Machine Learning and Artificial Intelligence Center at Severstal Digital, explained why this may happen. "There are people who say that they have worked for 20-40 years and know what to do better than an advisor. Sometimes this turns out to be true. An advisor needs to be modernized and improved continuously," Andrey Golov explained.
There are also cases when an employee who is busy with a lot of work does not have time to set up the advisor's work mode, so the company does not receive the effect of the models.
The third reason is that employees are not ready to work with the technology - they want to see a high-quality autopilot and will not settle for anything less. "If they see a simple cruise control, they refuse to use it. Therefore, the model has no effect if we do not switch to an autopilot," explained Andrey Golov, adding that, nevertheless, the use of advisors is justified in some cases.
Some participants in the discussion said that it would be better to move on to autopilots right away. Andrey Golov said that at their enterprise, most decisions are models that make control actions with a short periodicity - a second, a minute. "All models are actively involved in the process. Advisors are dying out, we have few of them left. We are aiming only at autopilots. If it is not an autopilot, then employees stop using the decisions," Andrey Golov said.
However, the participants noted the problem of the high cost of autopilots, which are expensive to implement. "Advisors are a very good start for further development for the SUUTP (improved process control system - ComNews note )," said Danila Malina, head of production digitalization at Uralchem. "It is expensive to immediately install an autopilot."
The participants in the discussion agreed that one does not exclude the other and that both autopilot and digital advisors can be used at the enterprise.
"It is more effective to use both approaches. Where autopilot is possible, use it. Where it is difficult to justify the costs of implementing additional automation, where it is not worth it or there is an operator, use this option," said Emil Gasanov, software product development manager at Nauka LLC.
At the plenary session, Deputy Director for Digitalization of JSC MCC EuroChem Valery Cherepanov said that, despite the movement towards automation, people will be needed at enterprises. "People will focus on other tasks. Therefore, we mean that unmanned production is possible, but we are unlikely to get there. We need those who will develop, maintain, and develop these systems," said Valery Cherepanov.