This is the Ph.D. research plan and progress report on a project aiming to make a Design Science contribution that comprises a modeling method and a model-driven architectural proposition for operationalizing the Work Systems Framework (WSF) by means of knowledge engineering. WSF practitioners have traditionally used semi-structured document templates for work system briefings, while academic works often emphasized the requirement for granular machine readability and traceability of such descriptions. The goal of this work is to achieve this by model-driven streamlining between (a) a domain-specific modeling language (DSML) founded on the WSF conceptualization and (b) large language model (LLM) services informed by diagrammatic work system descriptions, using (c) RDF graphs as a mediator format. This represents a convergence of artifact-oriented research agendas emerging in recent literature – the synergy of conceptual modeling with LLMs, and the GraphRAG patterns for enhancing LLM response quality with the help of knowledge graphs. The research follows a standard Design Science process that also encapsulates a DSML engineering cycle employing the Agile Modeling Method Engineering (AMME) methodology. Current results include the diagramming environment implemented on the ADOxx metamodeling platform including model-to-RDF transformation scripts to satisfy competency and traceability requirements. The second half of the Ph.D. timeline focuses on refining and deploying an evaluation protocol to measure the LLM responses quality (informed by the WSF diagrams as opposed to WSF documents) and to evaluate modeling language qualities.