AI: No success without strategy

Find the ideal approach for your company

Artificial intelligence: hardly any other term has sparked such heated debates in the last few years. Here, promises and warnings, and hopes and fears go hand in hand. When it comes to dealing with this field of technology, there is no one correct decision, but there is definitely a wrong one: anyone who ignores this development will lose out. Caution is, however, required: investments in artificial intelligence (AI) can only be successful with an individual strategy. The consulting experts from encoway with extensive experience in AI will provide you with support in finding the right way forward for your company.

In times of ever-increasing competition, a slowing economy and demographic changes, there is growing pressure to optimise process costs, to get by with fewer specialists and to retain existing customers while attracting new ones. AI approaches can provide effective help in meeting these challenges. This also means that if you are the first to venture towards AI, you may well be able to successfully outperform the competition. The value proposition of AI is thus enormous. However, whether solutions based on this technology can really be an effective lever for your company depends on many factors.

Many years of AI experience

At encoway, we have been working with AI for more than 20 years, both in practice and in research and development. In fact, a special AI method provides the basis for our software solutions. In other words, we know what we are talking about. This is why our most important preliminary advice is: do not believe everything you hear about AI, and above all, make sure that you critically question what you are promised. The bigger the promise and the more standardised the supposed solution, the more suspicious you should be. Because there is not just one form of AI, and therefore there is not just one type of solution that you can use for your company. Always ask yourself: what does my company really need and what route will lead most effectively – i.e. with feasible financial and time expenditure – to the goal?

There is no just one form of AI, and therefore not just one type of solution that you can apply to your company.

What makes it even more difficult to separate the wheat from the chaff, and to quite literally differentiate between fantastic promises and realistic applications, is the fact that there is no generally accepted definition of AI. Before we address the question of how your company can approach AI in an effective way, it is helpful to have a common understanding of the term. I will base my further comments on the following view: AI refers to systems that simulate human abilities, or behaviour that requires some form of intelligence. This includes perception of the environment, processing this input and the appropriate action in the environment. Or pragmatically: all methods that are able to find an (optimal) solution in a large number of ways in a reasonable amount of time.

The term “large number” provides a first crucial hint. A simple example: customers are able to choose their own personal favourites from 100 products. Anyone who makes use of AI here, is basically using a sledgehammer to crack a nut. Even a product configurator would be excessive for this task; the right tool would be a simple selector. If, however, you work with a variant space of 1030 like the drive technology and automation specialist Lenze, you have more possible combinations of products on offer than there are stars in the universe. Here it is simply impossible to reach the required selection within a realistic period of time without any intelligent technical tools.

AI strategy as a basis for decision-making

When assessing the potential that AI could offer your company, it is important to take the right perspective. You should therefore first define the problem and then consider which method can be used to solve it.t. This can, but does not necessarily have to be AI. The basis for decision-making is a comprehensive strategy that clarifies not only the objectives for the respective target groups but also what each solution will cost and what expertise is required. Anyone who rushes towards AI without such a strategy will almost certainly lose their investment in this field.

As a general rule, the more often a demanding task is repeated and the corresponding data is available, the more interesting AI approaches become.

For geared motors or valves, AI-based configuration can be a good option because they have large variant spaces with numerous component dependencies. However, for cables with standard lengths or colours, a simple selection based on conventional filters would suffice.

Areas of application for AI

There are many possible applications, and these should be carefully examined, always on an individual basis. There are, however, specific fields in which AI is already generating excellent results and which currently promise the greatest benefit. These include the areas of service and maintenance, production and plant optimisation and general data analysis. Here are a few examples: nowadays, a photocopier is able to realise when it needs a spare part based on measurement data. This type of AI application is referred to as predictive maintenance. In the area of service and maintenance, it is also possible to work on the basis of customer data, for example via the exact tracking of who ordered what and when. Correspondingly, the procurement of goods and warehousing can be optimally adapted to this. Configuration processes can be optimised on the basis of CPQ data combined with observed customer behaviour. Predictive models of this type result in improved production and capacity utilisation. Many companies already use collaborative robots or tools that are based on virtual reality or augmented reality, for example to allow unskilled workers to perform demanding tasks. The added value is obvious: the quality of the results, for example their accuracy, can be enhanced. Cost-intensive or manual – and sometimes dangerous – activities can be fully or partially automated. Large amounts of input can be pre-classified, errors can be analysed and people can be supported in their work.

Basically, there are no limits to the application areas for AI and new developments are broadening the horizon every day. Nevertheless, carefully weighing up costs and benefits is a key corporate tool. And this is precisely where the encoway consulting team comes in.. And this is precisely where the encoway consulting team comes in. Our advice draws on many years of experience in the marketing of modular products and successful variant management in machine and component construction as well as in related industries. This is accompanied by in-depth expertise in processes and data flows as well as a comprehensive understanding of the current challenges in the relevant markets. We employ this valuable knowledge as well as our extensive AI expertise in our consulting. On this basis, we provide advice for companies to help them identify their best individual AI strategy. If required, we can accompany this process with basic AI training to provide you with a general overview of this field of technology. As soon as we have defined the ideal AI strategy for your company together with you, the next step is to determine the specific problems to be addressed with AI. Finally, we will help you to select the appropriate methods and providers in order to effectively develop the defined fields of action with AI.

What can artificial intelligence do for your company? Is an AI approach advisable and how should you proceed? Please feel free to contact us; we will be happy to provide you with advice – without prejudging the outcome and independent of any specific solution!

Dr. Frank Dylla
Dr. Frank Dylla

Senior Consultant, encoway GmbH

Dr. Frank Dylla has been working on AI since 1997 and received his doctorate in the field of cognitive systems. He has gained experience in various areas of application, for example the control of autonomous robots, activity detection in video images, recognition and processing of natural language, architectural analysis and the efficient display of medical image information.