Are we entering the AI-driven technology age?

The event is being hosted in conjunction with Warwick Manufacturing Group (WMG) and will feature practical examples of applications of AI and data science for manufacturers, with insight from Volkswagen, Avery Dennison, Bitburger and more.

Ahead of the event, we caught up with Dr. Martin Liebig, Senior Director Data Solutions at Altair, to discuss how manufacturers are adopting AI projects to improve their processes.

How do companies bring AI applications into the factory floor?

Dr. Liebig: I basically see three approaches here. First, companies hire a team of data scientists. These then – as a central unit or data lab in IT – solve use cases as an internal service provider. In medium-sized companies, this approach is usually not an option because the investment is too high.

In addition, there is a lack of skilled workers. When we talk about specific solutions in manufacturing, we not only need experts who understand AI algorithms, but who can also answer questions about electroplating, for example.

Second, companies commission a classic service provider, such as a consulting firm, to solve use cases.

Third, democratisation. This is about making data analysis widespread. This is Altair’s solution approach, empowering companies across the board to create value from data.

We are a company with 30 years of engineering know-how. We draw on in-depth specialist knowledge that we have developed over decades of intensive collaboration with companies on specific applications. This broad engineering culture sets us apart as Our Center of Excellence approach combines data expertise with engineering expertise and we create a clear transfer of knowledge into the company and establish a natural business process.

The Altair RapidMiner Academy is a self-service learning portal that is open to everyone. When it comes to artificial intelligence, this format is particularly suitable for new customers. The biggest challenge for them in connection with AI is to apply the methodology to their specific challenges and to fundamentally understand how to use AI successfully. What needs to be done, when and with whom is explained quickly and easily in ten-minute blocks.

Following this introductory phase, this introductory phase, we move to small groups with 1:1 support. In this model, two of us work on a specific use case and teach the skills that our customers have learned in the Altair RapidMiner Academy to implement in their company and create value from them.

Which use cases will be implemented first?

Dr. Liebig: That is indeed the crucial question. If I come to a company as a consultant, we can easily uncover twenty use cases that could benefit from the use of AI in the factory . 20 cases that have different feasibility at different levels such as How difficult is the problem to solve? Is there current and precise data? Is there any data at all? How difficult is it to bring the results into an operational process later?

Companies are therefore well advised to carry out not only a feasibility study, but also a value creation study before deciding on the use of AI. For example, if it’s about reducing scrap, then it can be a very valuable ‘use case’ because companies can reduce their error rate – even if it’s only by half a percent. After all, this 0.5 percent contributes to the overall result.

Once the question of existing use cases and value-adding use cases has been clarified, we start a workshop where we clarify topics with the customer, such as is there a secure database connection? Where is the software installed so that enough computing power is available? Etc.

How long does it take to introduce AI processes?

Dr. Liebig: The goal for us consultants is to have models in production after a year. Experiencing AI applications live after just six months is a challenge, but doable. Twelve months must always be feasible.

To read similar articles, check out our Industrial Data & AI channel.