The AI Application and Test Center (ATKI) actively addresses the needs and challenges of SMEs. “The basic idea here is to support the entire development process at a company,” explains Professor Peter Schneider, one of the two heads of the Fraunhofer IIS Division for the Engineering of Adaptive Systems (EAS).

Depending on the company, ATKI addresses different challenges faced during the introduction of AI. Schneider identifies three specific challenges: economic benefit, technical feasibility, and the technology’s acceptance and maturity for application. The latter is referred to by the researchers as “readiness” and is the basic requirement for embarking on the AI journey. After all, companies find themselves at very different stages in terms of AI. When an SME approaches the team, it takes several steps to move from idea to implementation: A quick check to evaluate the prerequisites is followed by brief workshops lasting one day each.

The aim of these AI-readiness workshops is to evaluate a company-specific use case of AI, i.e., an idea for applying AI methods to optimize a process or increase the value of a product and whether it makes sense to use AI methods in implementation. As part of this one-day workshop, various aspects of the AI use case are queried, discussed, and documented. The key topics are:

• analysis of the process or system
• clarification of the availability and quality of data
• compilation of the people or divisions involved (i.e., those that are relevant to the use case)
• assessment of the current IT situation

An analysis of these topics paves the way for a realistic comparison of objectives and corporate reality regarding the AI use case in collaboration with the AI experts from Fraunhofer IIS. This AI maturity evaluation of the client’s selected use case takes account of economic, technological, and organizational aspects. In addition, a qualitative cost-benefit analysis is established for the use case.

Finally, individual recommendations are derived for the dimensions “Process/system,” “Data,” “Culture/mindset,” and “IT infrastructure” – and the next concrete steps toward implementation are set out.

SMEs with additional needs discuss their specific use cases in a consultancy process lasting up to 10 days and then enter long-term project guidance, which can last half a year. “This step-by-step approach allows us to address individual challenges effectively while remaining in constant, constructive dialog with companies,” says Schneider.

Various use cases already implemented

ATKI’s team consists of 35 people who have already implemented multiple projects with SMEs from Saxony and across Germany. Schneider recalls a project that implemented AI-based condition monitoring for axial pumps such as those used in construction machinery, shipbuilding, and process automation. “In such areas, system downtime is extremely costly,” he says. The identified solution will allow the error rate to be reduced significantly in the future. In another project, electrical currents were evaluated on synchronous motors in order to derive information on wear and the failures this may cause. The project was ultimately incorporated into product testing.

As well as condition monitoring, SMEs benefit from the technology in many other use cases, ranging from quality optimization to inline process control and worker assistance.