Artificial intelligence (AI) vs increasing the company’s efficiency

Algorithms that use artificial intelligence (AI) find wide application in the industrial sector. Thanks to them, we can confidently talk about the shift of production practices towards the 4th industrial revolution. Big words, isn’t it? We pose our traditional question here: how does artificial intelligence work in practice?

Artificial intelligence: dates, numbers, research

We have heard more than once that adopting strategies based on artificial intelligence (AI) improves the efficiency of the entire enterprise. So let’s look at an example of world economies and their attitudes to these issues. We know that:

• in 2010. Germany introduced the assumptions of Industry 4.0 as a development plan for its economy
• in 2011 USA has adopted the concept of smart manufacturing as part of the “Smart Manufacturing Leadership Coalition”[1]
• from 2018 France has developed a strategy for building a competitive advantage of its own economy based on artificial intelligence (AI) focusing on 4 areas: health, transport, defense and the environment[2]
• China, announcing the “Made in China 2025” strategy, emphasized the role of promoting advanced production while striving to become a leader in selected areas of artificial intelligence (AI)[3]
• Gartner predicts that by 2025. half of the organizations worldwide will reach the “stabilization stage” – artificial intelligence (AI) maturity[4]

Below is a map of the countries that have published a national strategy for the development and use of artificial intelligence (AI) in the economy:[5]:

How does it look in Poland?

In a study conducted by PSI in 2019. artificial intelligence solutions were considered important only by 21% of large enterprises (over 250 employees) and 27% of companies employing 50-249 employees[6]. As you can see, artificial intelligence was not considered one of the basic strategic activities of Polish enterprises at that time. The same study strongly indicated that investment in advanced IT systems for production is of key importance for companies. Most importantly, this issue has been recognized as one of the most commonly used elements of the assumptions of Industry 4.0 in Poland. In this way, digitization and IT modernization were introduced to our economy. The first stage had to be an investment in IT systems. What’s next?

Now the situation has changed considerably. According to the data presented in the “State of Polish AI 2021” report[7] our country is ranked 1st among CEE countries in terms of experts working on artificial intelligence. This result shows how much the approach of Polish enterprises to introducing IT modernization has changed. Currently, in order to develop a competitive advantage, it is not enough to collect valuable data. First of all, the possibilities of artificial intelligence (AI) should be used for efficient data management, their automatic selection and drawing conclusions. The information processed in this way helps the management to make effective business decisions. The analysis of the huge amounts of data that flows into enterprises from various sources must lead to the presentation of specific conclusions. Such approach enables quick reaction to dynamic economic changes and adaptation of production to the expectations of customers.

Artificial intelligence (AI) in practice

Theoretical considerations regarding the impact of artificial intelligence (AI) on the industry have resulted in a series of studies in those companies that have implemented these assumptions in practice. The results are extremely promising. 82% of manufacturing and service companies admit that their efficiency has increased thanks to the use of AI.[8] These assumptions have a positive impact on three levels of each enterprise:

• system
• process
• product

Real-time process monitoring is possible thanks to the combination of two elements: intelligent sensors installed directly on production machines and advanced computing algorithms (AI, ML, DL). Machines connected by a wireless network send data to one system. This information is used to support decision-making processes, increase the effectiveness of activities and facilitate work.

Examples:

• effective planning and implementation of corrective and maintenance activities
• optimization of services and controlling the inventory of spare parts
• advanced control of the entire product life cycle (from raw material extraction to production and logistics processes)
• real-time status monitoring of individual processes, quality control
• forecasting the demand, shortening the time of the production cycle, waste management, etc.
• ensuring the highest quality of the product

 Why is it so difficult to make good business decisions today?

Nowadays, manufacturing activity is based on technologies that use data. Huge amounts of data. Most companies keep their own registers that allow for a very basic systematisation of this information. However, the number of possible channels for the flow of valuable data can be overwhelming. Social media, loyalty systems, customer purchase registers, customer communication registers (chats, telephones, e-mails, mobile applications), complaint registers, or customer behavior patterns with specific attributes – there are many examples. A responsible approach to business involves making key decisions based on the current knowledge about the condition of the brand and its customers. How can you get it?

Entrepreneurs’ expectations are clear: integrating data in systems that will be able to suggest optimal solutions. The new approach to digitization focuses on these aspects. It is no longer enough to collect data. Now it is necessary to use the achievements of IT to make decisions on a large scale. All with real data analyzed in real time. In practice, algorithms based on artificial intelligence (AI) make it possible to implement these demands.

Over the past few years, the manufacturing sector has clearly seen an increase in interest in terms such as machine learning (ML) and deep learning (DL). All because of the ever-expanding datasets. Companies have to analyze them and draw specific conclusions based on them. Extracting the most valuable data is possible thanks to deep learning (DL) algorithms. These algorithms are based on artificial intelligence (AI) and machine learning (ML). All these elements are intended to imitate the way in which a person acquires certain knowledge.

 Artificial intelligence and predictive maintenance

Predictive maintenance is one of the basic tools used to generate real savings in enterprises. The larger the scale of the company’s operations, the greater the effects of predictive maintenance. This approach in the context of the assumptions of Industry 4.0 includes planning the maintenance of products or individual devices that are equipped with the machine park in the company. The service should be carried out at the most optimal time to reduce possible downtime and fully use the potential of the parts operating in the machine. How do we know when this optimal time is coming? Artificial intelligence (AI) provides the answer.

Predicting failures in manufacturing enterprises using intelligent sensors and the data collected from them helps to shorten / eliminate downtime of entire production lines. The use of machine learning (ML) algorithms makes it possible to combine data collected in real time with historical information. Based on this, the system draws conclusions about maintenance. But that’s not all! On the basis of predetermined parameters (e.g. the manufacturer’s forecast of wear of individual mechanical parts used in production machines), the system informs about which elements need to be replaced as soon as possible and which are the end of their service life. This allows you to plan effective maintenance in advance, but also provide access to the necessary spare parts without having to overburden internal warehouses.

Optimal patterns of work are generated by systems even from a small amount of data. On this basis, you can generate automatic prompts supporting the work of the entire production. The more information provided to the IT system, the more accurate these automatic recommendations will be.

Focus on quality

Considerations about the role of artificial intelligence (AI) in the process of increasing the efficiency of manufacturing enterprises should also apply to advanced quality control. After all, providing the best end product is the basic postulate of every company. Artificial intelligence (AI) and deep learning (DL) open up great opportunities in this context. Product data is computer-processed to detect defects even in the early stages of the production process. The margin of error is minimal here. A defect identified early enough often allows to carry out repair processes so that the final version of the product does not deviate from the adopted standards. This approach generates significantly greater savings than reworking a non-compliant end product.

Example? Research carried out in a factory producing clothes shows that the use of AI and DL technologies allows to achieve the detection of stitch defects at the level of over 92%[9]. Artificial intelligence has a positive impact not only on the quality of products, but also on work safety and production efficiency.

Conclusion: adopting an approach based on AI, DL, ML algorithms allows to better predict and control the quality of production. Additionally, companies can make business decisions based on much larger data sets obtained from various sources. These data are processed instantly.

Why is it so important? Artificial intelligence (AI) is used in almost every aspect of the functioning of production companies: from predicting possible failures, through optimal resource management, to advanced quality control and supporting sustainable development.

Is it worth investing in AI? It depends on how much you want to gain an advantage over your competitors😉

---

[1] A.Jamwala, R.Agrawala, M.Sharmaab, Deep learning for manufacturing sustainability: Models, applications in Industry 4.0 and implications, International Journal of Information Management Data Insights, Volume 2, Issue 2, November 2022, 100107 https://www.sciencedirect.com/science/article/pii/S2667096822000507?via%3Dihub

[2] https://arp.pl/documents/42/Raport_ARP_2018_Przemysl_4_0.pdf

[3] C.G. Machado, M.P. Winroth, E.H.D. Ribeiro da Silva, Sustainable manufacturing in Industry 4.0: An emerging research agenda, International Journal of Production Research, 58 (5) (2020), pp. 1462-1484, 10.1080/00207543.2019.1652777

[4] https://www.gartner.com/en/newsroom/press-releases/2021-11-22-gartner-forecasts-worldwide-artificial-intelligence-software-market-to-reach-62-billion-in-2022

[5] https://fintechpoland.com/wp-content/uploads/2022/03/AI_raport_FIN-1.pdf

[6] https://przemysl-40.pl/index.php/2019/11/04/cztery-raporty-o-przemysle-4-0-w-polsce/

[7] https://www.erp-view.pl/it-solutions/30104-przemyslowe-ai-jak-sztuczna-inteligencja-rewolucjonizuje-przemysl-produkcyjny.html

[8] S. Akter, G. McCarthy, S. Sajib, K. Michael, Y.K. Dwivedi, J. D’Ambra, K.N. Shen, Algorithmic bias in data-driven innovation in the age of AI, International Journal of Information Management, 60 (2021), Article 102387

[9] A. Abou Tabl, A. Alkhateeb, W. ElMaraghy, Deep learning method based on big data for defects detection in manufacturing systems Industry 4.0, International Journal of Industry and Sustainable Development, 2 (1) (2021), pp. 1-14