Steinbeis experts co-develop an AI-driven support system to improve the management of skills requirements
Companies currently face a variety of complex challenges, with constant changes in the business environment compelling them to question their business models – in some cases forcing them to redesign them. This also underscores the importance of the process of developing the skills required by companies. But is there an effective way to design this process? This was the question looked at by a team at Innovation Engineering, the Steinbeis Innovation Center, and design specialist Denzinger. Together, they developed XPRT, an AI-based assistance and recommendation system that helps with the selection of individual learning content.
Companies, especially SMEs, currently face huge challenges, with war in Europe, complex relations with China, due diligence obligations in the supply chain, sustainability requirements, ESG, carbon neutrality, and even technological paradigm shifts such as AI and additive manufacturing. Often, companies find themselves having to completely rethink business models and value creation processes. This is exacerbated by skills shortages and, looming on the horizon, a generation change within the workforce. Companies often grapple with the best way to gear themselves to future needs and what their new core competencies should be. This development has unsettling consequences for Germany as a business location. Many SMEs are steering clear of the hard option, which would necessitate business transformation, and are thinking about giving up.
Organizing competencies better and making it easier to take action
In addition to economic and individual factors, the much-lamented shortage of skilled workers – or skills shortage – is always linked to company culture and the subjective satisfaction of individual employees in their different roles and positions. Among other things, this satisfaction also depends on whether people feel over- or underchallenged. In essence, when workers don’t feel comfortable, they leave. People come to a company and if they’re happy, they stay.
But it’s rare for these issues to be seen in the right context: At many SMEs, there’s a disconnect between strategic orientation, which determines future viability and required competencies, and the issue of training and thus skills development. Human resources and corporate learning are rarely considered tactical tools that can be used to realign the company. In keeping with the principle of making businesses future-proof, human resources should be managed just as professionally as technology and finances.
This problem is reflected in how companies gather and use data. Business intelligence is mostly about monitoring and networking machines, processes, or products. The company workforce, and thus competencies within the company, are generally left out of the equation, and that’s ignoring the issue of staff training and the value that measures bring to business practice.
The only potential way forward is to systematically manage skills on an integrated basis, also capturing future needs in ways that concur with strategic requirements and tactical measures – business competence – and this should place equal emphasis on the needs of both companies and employees, and bring those needs into harmony. Action can be taken by adopting two opposing approaches:
- Top-down approach
The skills requirements of the company are laid down by management and are derived from the business strategy and corporate vision. Requirements are context-dependent and different for each company: People working in product management at a medical technology startup require different skills from product managers at a medium-sized automotive supplier, and for a startup in medical technology, the emphasis of AI is different from a medium-sized automotive supplier. - Bottom-up approach:
The skill profiles of individual employees or entire teams capture, among other things, professional experience, education, interests, and personal characteristics. This makes it possible to contrast profiles with the needs of a company, to point to existing competencies, and to highlight any areas in which further training is required – or the skills new hires should bring to the company.
Linking business goals to the right skills
As part of the Learn4U project funded by the Adolf Leuze Foundation, the experts at Innovation Engineering have developed a procedural model and an AI-based recommendation system that helps SMEs gear the business and staff better to future demands, also helping firms navigate their way through the jungle of continuing education topics. It quickly became apparent that much work is needed in this area, especially when it comes to current requirements, and that it is too much to expect training to satisfy needs if specifications are unclear. Currently, many learning technology startups draw on the benefits of matchmaking, but they often fail to take into account the contextual requirements that are crucial for business practice.
This contrasts with the project team comprising the Steinbeis Innovation Center and Denzinger, which believes in interlinking the strategic goals of organizations with corresponding tools of skills development in HR, in keeping with the V-model or V-model XT used in systems engineering. The model developed by the joint team, the aim of which is to identify the skills requirements of individual clients, should help firms safeguard future viability: Drawing on a variety of methods of requirements engineering, human-centered design, and systems engineering, this model makes it easier to capture and assimilate ambiguities in the business environment, volatile customer needs, and the corporate culture within the company.
Regarding employees, the system logs professional know-how (skills) and the kind of soft skills and personality traits (abilities) that are central to many roles. It also includes market and industry information in order to match individual needs to the business environment. This makes it possible to paint an accurate picture of necessary learning needs: Which specific skills will I need at my company in the future and which not? And accordingly, who will need to learn what? The second step is to compare and contrast these requirements with current offerings. Requirements can also be returned to in the future to gauge success.
The procedural model developed for the project comprises the following modules:
- Gap analyses:
Formal collation and evaluation of (company- and project-related) skills requirements and the (employee-specific) availability of skills - Skills development plans:
Formal collation and evaluation of made-to-measure learning content and formats for use in internal and external curricula - Personalized upskilling and learning processes:
Formal design of individual and group-specific training schedules or learning paths with an assistance and monitoring function
XPRT: structured, transparent, efficient
Early on in the project, it became evident that the process could be supported by an intelligent, AI-driven assistance and recommendation system. This system would offer a variety of benefits, such as efficient processing and usability, help with capturing and inputting information, more structured output, reproducible methods, and improvements in the quality of results thanks to scalability (broader datasets).
The XPRT recommendation system, which was subsequently developed iteratively with the Institute for Business Analytics at the University of Ulm, provides assistance with a variety of tasks and offers the following example features:
- Input of master data from companies: such as PDFs containing descriptions of strategies, which XPRT uses to automatically derive skill requirements.
- Input of employee know-how: such as resumes from LinkedIn profiles, which XPRT uses to automatically derive the skills and know-how of employees.
- Output: recommendation of suitable people to work in defined (project) roles.
- Output: comparisons between self-evaluations and peer assessments regarding competencies.
- Output: course recommendations, with priorities, for individually developing skills, including an explanation of why particular courses should be considered.
Using so-called explainable AI makes it possible to gain maximum transparency regarding the recommendations made by the system. Users are free to dismiss recommendations at any time and tweak underlying inputs as and when desired (“human in the loop” approach). It is proven that this raises the acceptance levels of learning recommendations, making users more likely to want to follow those recommendations.
Skills – the currency of the future
It is often said that data is the new gold. In keeping with that metaphor, skills are the new platinum. Given the right processes and an intelligent tool, this precious metal – so important to the SME sector and its future – can be mined and refined. The major benefit of the XPRT method is that it offers ultimate flexibility in interlinking a variety of sources of company information (it can even draw on unstructured data), thus systematically dovetailing staff, such an important resource of any company, with the business strategy. XPRT simplifies and systematizes the processes this requires and in doing so, it offers the possibility to do much more than simply identify suitable training. For example, in the future XPRT is expected to make it possible to
- support management with the alignment of corporate strategies to market environments
- simplify training processes by clearly identifying required skills during recruitment, and enhance the efficiency of learning (ROI), thus finding exactly the right people to fill vacancies
- identify gaps in HR management between self-assessments conducted by employees and the assessments of their seniors, thus making it possible to address business-critical skill requirements through individualized measures
Only reasonable effort is required – of companies or their workers – to introduce the XPRT system. Existing documentation on areas such as the business strategy, projects, but also qualification profiles, resumes, and personal performance goals, can be adapted to make them machine-readable so they can be entered directly into the tool. Linking the system to different learning/education options (internal and external) allows XPRT to quickly generate initial up-skilling and re-skilling recommendations for selected employees.
The results described above were achieved as part of the Learn4U project and supported in financial terms and in spirit through funding provided by the Adolf Leuze Foundation in keeping with its goal of promoting training, education, science, and research.