“Successful AI adoption is not about chasing the newest model; it is about improving a clearly defined business process responsibly, one manageable step at a time.”
There is no single route towards AI adoption. Every region, organisation and SME starts from a different position. In our Meet the Boost AI Partners series, we introduce the people and organisations working together to make Artificial Intelligence more practical, safe and accessible for smaller businesses across North-West Europe.
This time, we meet Steffen Reithermann, AI Project Manager at the Machine Learning Group at RPTU University Kaiserslautern-Landau and Managing Director of the AI Project Office of the AI Alliance Rhineland-Palatinate.
Why should SMEs start exploring AI now?
For Steffen, AI is no longer a specialist technology available only to large corporations. It is increasingly embedded in standard business software, cloud services and affordable tools, allowing SMEs to begin with focused experiments rather than major transformation programmes.
At the same time, customers, competitors and employees are already changing how they work. Waiting can mean losing valuable learning time or allowing informal AI use to develop without clear rules. SMEs therefore need practical AI literacy: an understanding of what AI can and cannot do, where it can create measurable value and which data, security and compliance questions need attention.
Exploring AI does not mean adopting every new tool. It means gaining enough experience to make informed decisions, choose suitable use cases and remain in control of the company’s digital development.
“Where should we begin?”
When SMEs approach RPTU about Artificial Intelligence, the question Steffen hears most often is: “Where should we begin?”
This is followed by practical concerns. Is the company’s data good enough? Which use case could save time or improve quality? Can employees enter company information into a public AI service? How do GDPR, the EU AI Act, cybersecurity and intellectual property apply? SMEs also ask about costs, dependence on individual providers and whether AI will replace jobs.
Steffen also sees two common misconceptions: that useful AI always requires a large custom model and a dedicated data-science team, or that a generative AI tool can simply be switched on like ordinary office software. For RPTU, the task is to “separate opportunity from hype” and turn these concerns into manageable business, technical and organisational decisions.
Creating a trusted starting point
If Steffen could remove one barrier to AI adoption, it would be the absence of a trusted and accessible starting point. SMEs face vendor promises, technical terminology, rapidly changing products and uncertain legal or security implications. With limited time, this complexity can lead either to paralysis or unstructured experimentation.
“The barrier is therefore not only a lack of technology or funding; it is a lack of orientation.”
A company needs a neutral pathway that connects a concrete business objective with the relevant process, data, people, risks and expected benefits. The next step should then be small and testable rather than a major commitment. This is why self-assessment, AI literacy, risk and opportunity tools and support from trusted regional organisations are important within Boost AI.
Connecting research, regions and businesses
RPTU brings a combination of current machine-learning research and a multi-level transfer network to Boost AI. At European level, Boost AI connects RPTU with partners across six countries. At state level, the AI Alliance Rhineland-Palatinate links universities and universities of applied sciences, while its AI Project Office at RPTU supports visibility, exchange and cooperation. Regionally, SIAK connects research organisations and innovative companies in and around Kaiserslautern.
This creates a feedback loop: emerging technologies, risks and research findings can be translated into practical guidance, while questions from SMEs can inform tools, training and future research. RPTU also contributes AI horizon scanning and critical technology assessment and can draw on Kaiserslautern’s wider AI ecosystem with many prominent affiliated research institutes.
Start with the business process, not the technology
“Making AI practical means starting with a business process, not with a model or fashionable product.”
For Steffen, practical support should help an SME answer a few concrete questions: What problem are we trying to solve? Who benefits? Which data are involved? What could go wrong? How will we measure whether the solution works? And what is the next realistic step?
Accessibility requires plain language, short modules, relatable examples, templates and self-assessments rather than long technical reports. Human support is also important when a company reaches a difficult decision. For an SME with limited time and budget, a good first step may be improving one repetitive workflow with functionality already available in existing software.
Employees should be involved early, the pilot should be low-risk and success criteria should be defined before scaling. As Steffen puts it: “Practical AI creates clarity and learning, not additional complexity.”
Shared questions across different regions
Working with partners from six countries has shown Steffen that regional contexts differ, but the core questions from SMEs are remarkably similar. Support structures, priority sectors, digital maturity and available expertise vary, yet businesses repeatedly ask where to start, how to identify value, how to protect data and how to involve employees.
For Steffen, this confirms that Boost AI needs a shared foundation without becoming a one-size-fits-all programme. Tools should be transferable across regions while remaining flexible enough for local sectors, languages and support systems. He also sees value in the complementary perspectives of research organisations, public authorities, business support organisations and innovation networks.
What would success look like in 2028?
For Steffen, success in 2028 would mean more than reaching project targets. SMEs should have the confidence and capability to evaluate AI independently: identifying a useful case, assessing providers, managing data and risks, involving employees, testing a solution and measuring its value. Just as importantly, they should know when not to use AI.
Business Support Organisations should have trained staff, proven methods and a living toolkit that can be adapted to different sectors and levels of maturity. Local and regional authorities should have embedded AI more clearly in their digital and innovation strategies, while the transnational partnership continues to update tools, share evidence and learn from implementation.
Ultimately, success means that smaller businesses outside major innovation hubs are no longer excluded from responsible AI adoption.
From single answers to multi-step AI assistants
Outside Boost AI, Steffen is particularly interested in the combination of agentic AI and smaller, domain-adapted models. AI is moving from systems that generate a single answer towards assistants that can plan, retrieve knowledge, use tools and support multi-step workflows.
Combined with local or private deployment, clear guardrails, logging and human approval, this could give SMEs powerful capabilities while retaining greater control over sensitive information and business processes. But Steffen stresses that “impressive demonstrations are not the same as reliable operations.” Security, error handling, evaluation and responsibility become even more important when AI systems can take actions.
One thing every SME should understand about AI
Asked what he hopes every SME in Europe will understand about Artificial Intelligence, Steffen returns to the importance of starting with the business rather than the technology:
“Successful adoption is not about chasing the newest model; it is about improving a clearly defined business process responsibly, one manageable step at a time.”
Steffen believes Boost AI is relevant not only to SME owners and managers, but also to employees, Business Support Organisations, authorities, technology providers and researchers. SMEs can gain practical starting points, examples and tools, while support organisations and authorities can learn what smaller businesses need in practice.
The project is particularly relevant to organisations outside major technology hubs and to anyone interested in strengthening competitiveness without lowering standards for data protection, security, transparency and human responsibility.
About RPTU University Kaiserslautern-Landau
RPTU University Kaiserslautern-Landau is the technical university of the German state Rhineland-Palatinate with a leading computer science department. Its main contribution to Boost AI comes from its Machine Learning Group, which combines internationally renowned research in machine learning and deep learning with application-oriented projects, knowledge transfer and AI consulting for companies.
Within Boost AI, RPTU contributes current machine-learning research, its multi-level transfer network, AI horizon scanning and critical technology assessment.
About Boost AI
Boost AI brings together partners across North-West Europe to help smaller SMEs build their capacity to use Artificial Intelligence in a practical, safe and responsible way. The project focuses particularly on SMEs in agrifood, hospitality and small-scale manufacturing.
Follow Boost AI for more partner stories, project updates and practical insights into AI adoption for SMEs.