In the immersive complexity of digital innovation, data-driven solution design often begins with a specific technology or potential solution already in mind. However, this predisposition narrows the design space and reduces the probability of discovering more suitable or novel alternatives. This thesis introduces a formally developed ontology to support datadriven solution design, offering a meta-structure that facilitates abstraction and translation from business opportunities to technological concepts. Developed through extensive literature reviews and four empirical case studies, the ontology was constructed using both deductive and inductive methods. It uniquely integrates the layers of Business, System, Data, and Technology, enabling structured reasoning and navigation across abstraction levels. The hierarchical ontology serves as the foundation for a multi-agent large language model system (MALLMs) that assists users in identifying appropriate solution concepts aligned with provided business opportunities. Central to this system is a Knowledge Graph (KG) acting as a meta-agent, providing structural guidance and validation. Additional agents generate solution-relevant information through retrieval-augmented generation (RAG) based on a curated corpus of over 14,000 research papers, academic articles, and code fragments related to data-driven solution design. The findings suggest that combining symbolic representations with neural models enhances output effectiveness by leveraging both user and domain-specific knowledge.