AI in Mechanical Engineering: Why Companies Have Nothing to Fear

The launch of ChatGPT at the end of 2022 has fueled a discussion among the general public that has been going on in the industry for years: How useful or dangerous is artificial intelligence (AI) and what role will human decisions still play in the future? 

Industrial companies have not just been concerned with the potential of machine learning technologies (ML) and algorithms since ChatGPT: An IDG study shows that 73 percent of large companies with more than 10,000 employees have already worked with ML by 2021. According to a recent Ifaa study, 65 percent of manufacturing companies with fewer than 50 employees are not considering using AI.  

Nevertheless, many executives feel pressured by the AI trend. Sixty-seven percent believe they need to adopt AI to stay competitive, according to a recent AI study by AI vendor Monolith. In fact, industrial AI users estimate they are 43 percent more likely to see increased revenue and profitability compared to non-users — and apparently in record time: in the IDG study, the majority of ML users reported benefiting from the technology in three months or less.  

Put simply, AI is the ability of computers to learn, to recognize patterns, and to make decisions based on them. This involves the use of various technologies such as machine learning, natural language processing, and computer vision. However, companies in the engineering sector that are looking at the possibilities of AI for their everyday operations should also know what AI is not. 

AI tools are not an out-of-the-box solution that solves all your problems and meets your requirements 100 percent. Rather, AI is a toolbox: companies can make use of it and develop individual solutions that support people as best as possible in their everyday work.  

The manifestations of this are manifold: AI processes can  

• run in the background and perform tasks such as data preparation and cleaning, or 
• interact directly with the user — this is the case, for example, with voice assistance systems, chatbots, and autonomous systems such as robotics applications. 

AI has long been an integral part of our everyday professional and private lives: facial recognition in smartphones, various online services, driver assistance systems and voice assistants are just a few examples of AI applications that have become part of life.  

AI is also leaving its mark on industry in areas such as predictive maintenance, smart farming and wherever the analysis of data plays a role and intelligent assistance systems take work away from people.  

In principle, AI helps to make applications more intelligent. This is especially important against the backdrop of demographic change, which is causing more and more knowledge to retire. Companies are often already very good at collecting and distributing information. However, AI helps to interpret this information and derive specific recommendations for action to meet the requirements of different users. In this way, AI counteracts the loss of knowledge. 

Concrete use cases for AI in mechanical engineering can also be found in aftersales & service as well as technical documentation. Typical tasks that artificial intelligence can support there are: 

• Spare part recognition 
• Predictive maintenance  
• Identification of upcoming maintenance needs 
• Quality assurance and machine translation of technical documentation 
• Customer self-service portals 
• Chatbots for aftersales support 
• Networking of service-relevant information 

The following examples show the added value AI offers for aftersales & service and technical writers. 

The basis for AI to generate added value for users and customers is well-prepared and intelligently networked data. At Quanos, for example, this data serves to simplify manual processes and save users’ time. One example is the intelligent preparation and provision of data from PDF files in digital form.  

Service organizations generate a lot of valuable data over time. However, historical data sets are often only available as scanned PDFs. AI-powered tools help to easily digitize the information: The tools can learn to extract specific information selectively and recognize contexts and elements such as tables, paragraphs, or forms. In this way, AI tools offer efficiency advantages over optical character recognition (OCR).  

Metadata, which is becoming increasingly important for networking and distributing data, can also be enriched with the help of AI. One of the strengths of AI is its ability to classify content at lightning speed. AI is capable of evaluating large data sets in a short time and recognizing patterns as well as anomalies.  

To ensure high data quality, a confidence value is often used for the interpretation of AI results. This value indicates how certain the decision of an AI model is. Companies can set a threshold and manually check results when they fall below this value. This approach requires manual reworking. Nevertheless, users gain efficiency because they do not have to start from scratch.  

Technical writers also achieve a gain in efficiency if they have sections of text created by a Large Language Model (LLM) — a generative language model with artificial intelligence — and then correct them as an "expert-in-the-loop" and give them the finishing touches. 

Although data preparation is time-consuming, the investment is worth it. The better the data is prepared and enriched with metadata; the better users can work with it using AI. Information about machines, spare parts and processes is one of the most important tools in aftersales & service. 

AI helps service technicians find the data they need to process service calls, even in a flood of information. Artificial intelligence is able to deliver the right result even when there are typing errors and to make contextual, intelligent suggestions that speed up the calls’ resolution.  

Whether it's processing a foreign-language search query, finding spare parts by photo, or being controlled by voice: AI can substantially simplify access to information in service and significantly increase productivity. The prerequisite for this is to always keep the needs of the technicians in mind.  

At Quanos, many of the possibilities offered by AI in mechanical engineering can already be put to practical use today. For example: 

• Content created with the XML editing system ST4 can be machine translated with ST4 Smart Extension AITranslator using AI methods.  
• The ST4 AI Jetpack independently enriches metadata to simplify data preparation.  
• AI Search helps service technicians find information. 

In this blog post, we will take a look at what questions mechanical engineering companies that want to use AI should ask themselves and in this blog post you will find handy tips for the practical implementation of AI in mechanical engineering.