Designing data for reuse before its second life begins
In many industrial and regulated environments, data is originally produced for a primary operational purpose such as process control, quality assurance, or compliance. The growing demand for AI, advanced analytics, and optimisation requires that the same data can later be reused for secondary purposes that were not part of the original system design. In practice, this secondary use is often difficult: data may be high-quality in its original context but lacks the semantic structure, documentation, and contextual meaning needed for reuse across departments, analytical teams, or AI-driven workflows.
The SDR challenge concerns how data can be made semantically reusable beyond its initial intended purpose. Semantic readiness is not only about cleanliness or technical accessibility, but about whether the meaning of data remains stable, interpretable, and transferable when applied in new contexts. The case stresses early semantic design as a long-term enabler: building SDR into data flows from the beginning rather than retrofitting semantics after systems are in production.
How does Semantic Data Readiness influence the ability to reuse operational data for secondary AI and analytics applications?
Data designed with explicit semantic grounding from the outset significantly increases its long-term reuse potential and reduces the effort required to scale AI beyond isolated PoCs.
Datasets representing typical primary-use industrial data streams will be evaluated under different SDR maturity levels. Comparative experiments will explore how semantic enrichment affects downstream AI performance, interpretability, and integration effort when data is repurposed for secondary analysis. Scenarios may include quality-controlled process data reused for predictive models, knowledge extraction, or decision support.
Practical exploration of how semantic structure can be introduced into existing regulated data landscapes while maintaining traceability, validation requirements, and compliance. The case provides insight into organisational barriers and enabling strategies for moving from operational data silos to semantically reusable knowledge assets.
Links closely to AI-agent cases (UC001) and governance-oriented cases (UC007, UC008), since secondary use often requires shared terminology, metadata, and cross-domain semantic alignment. Supports the broader project ambition of reducing the number of AI initiatives that stall at PoC level.