This work used Design Science Research Methodology to investigate the enterprise AI gap through problem framing, domain-grounded inquiry, artefact creation, and structured demonstration.
The goal was not only to understand the problem, but to build through it.
The study moved from enterprise AI tension, into a regulated power transmission context, and then into two contrasting build paths.
The problem could not be resolved through abstract policy analysis alone, nor through a loose prototype exercise. It needed artefact-based inquiry that could make governance constraints, design assumptions, and operational trade-offs visible inside a working system.
DSRM provided that structure: a disciplined way to frame the problem, ground it in a real domain, build deliberately, and examine the resulting systems credibly.
How can organisations safely adopt Generative AI when its inherently non-deterministic behaviour is in direct tension with the governed, rule-bound, and auditable structures that enterprises must maintain?
This question framed the entire study. It shaped what needed to be understood, what had to be built, and what would count as a meaningful research contribution.
The research began by defining the core tension between probabilistic AI behaviour and deterministic enterprise governance. This was not treated as a minor implementation issue, but as a structural mismatch at the heart of enterprise AI adoption.
The next step was to define what a safe enterprise AI solution would actually need to achieve. That meant making the incompatibility gap explicit, defining acceptance expectations, identifying design requirements, and recognising that any workable approach had to support both transformative experimentation and governed operational improvement.
From that base, the research moved into building. The problem was approached from two directions at once: a transformative direction that explored new possibilities, and an incremental direction that worked within existing governance and operational constraints.
These two directions became the basis for the two systems developed in the study.
The research did not search for a single generic AI answer. Instead, it deliberately moved into design and development through two contrasting systems, each built to surface a different side of the enterprise AI challenge.
GridSense explored the open, transformative side of enterprise AI. It was developed rapidly and experimentally to test how AI could help create a new kind of operator experience rather than simply automate an old one. It represented the space where ideas are discovered, interfaces are reimagined, and AI is used as a creative development partner.
OsciProbity explored the structured, auditable side of enterprise AI. It was shaped by requirements for governance, traceability, confidence-aware logic, and architectural discipline. Its development emphasised specification, design review, and AI-assisted implementation within a more controlled engineering process.
The next page begins with these two systems and what each was designed to test.