Client Overview
The UK Nuclear Decommissioning Authority (NDA) oversees one of the largest and most regulated project portfolios in Europe — spanning complex, safety-critical, multi-billion-pound programmes.
To prepare for next-generation oversight and predictive decision-support, the NDA commissioned Surtori to define the Data Standards and Governance Framework that would underpin its first AI-enabled proof of concept — SATNAV (Situational Awareness, Tracking, Navigation and Verification).
The initiative aimed to shift project controls from reactive reporting toward predictive insight and proactive optimisation, establishing the data foundations for AI and machine-learning capability across all operating companies.
The Challenge
For the NDA, the obstacle wasn’t a lack of systems or data — it was a lack of trust.
Over decades of decommissioning work, every operating company within the NDA group had built its own data practices, formats, and governance rules. Information was technically abundant but operationally incompatible. Performance reports from different sites could not be compared, aggregated, or modelled with confidence.
Adding to the complexity:
- Data originated from disparate contractor ecosystems, each with separate assurance chains.
- Safety and compliance obligations demanded auditability and lineage, not just dashboards.
- The organisation faced regulatory scrutiny that required any AI or predictive system to be explainable, traceable, and defensible.
- Legacy terminology and inconsistent metadata created semantic noise that undermined decision-makers’ confidence.
The core challenge, therefore, was not automation or reporting — it was establishing a unified data language across a federated enterprise where governance, risk, and technology maturity varied wildly.
Surtori’s task was to design the framework that could bridge these divides — a set of standards robust enough for nuclear-sector regulation, but flexible enough to enable AI adoption across a diverse ecosystem.
The Surtori Approach (Fusion Framework™)
Using Surtori’s Fusion Framework™, the engagement engineered the NDA’s first enterprise-wide data-governance and AI-readiness architecture, validated against DAMA-DMBOK, ISO 8000, and ISO/IEC 11179 standards.
| Phase | Focus | Key Actions | Outcome |
| Discover | Diagnose existing data environment | Conducted full audit of project-controls submissions and system interfaces across operating companies. | Identified 170+ data elements requiring harmonisation. |
| Design | Build standards and governance model | Authored NDA Data Standards and Data Contracts Framework, including catalogue, dictionary, ontology, and data-quality KPIs. | Established a controlled, interoperable data model forming the blueprint for AI use. |
| Deliver (POC) | Enable AI feasibility | Defined data pipelines, schema validation logic, and AI model-readiness scoring for the SATNAV POC. | Demonstrated feasibility of predictive variance and performance forecasting using NDA data. |
| Embed | Institutionalise good data practice | Designed a 4-phase roadmap — from foundations to predictive analytics — and trained internal teams on data stewardship. | Framework adopted by the NDA Chief Data & Analytics Office (CDAO) as reference architecture for future AI initiatives. |
The Results
Within 12 weeks, the NDA achieved measurable progress from manual reporting toward intelligent data control:
- Unified Data Dictionary: 400+ fields standardised across risk, cost, and schedule domains.
- Data Contracts Implemented: Quality thresholds, schema, and change-control rules enforced across operating companies.
- Predictive Readiness: AI models validated with >85% accuracy in test forecasting.
- Governance Framework Established: 7-step model now integrated into the NDA’s enterprise data strategy.
- Scalability Proven: POC confirmed a clear pathway from concept to enterprise-grade predictive system.
Strategic Insights
Analysis of the SATNAV case studies informed Surtori’s recommendations — drawing lessons from Lightricks, Iron Mountain, MLB, and NatWest Markets:
- Unified architecture enables single-source-of-truth governance.
- Automated data-quality scoring reduces human error and enhances trust.
- Self-service analytics accelerate cultural adoption.
- Regulatory alignment ensures auditability and public accountability.
These lessons validated that the NDA framework matched proven global success patterns, positioning it as one of the UK government’s first operational AI-readiness blueprints for project controls.
Legacy and Ongoing Impact
The NDA Data Standards Framework is now embedded across its wider digital-transformation roadmap, forming the data backbone for predictive assurance, portfolio analytics, and AI-governed decision-making.
Surtori continues to advise on data-governance scaling and model-performance validation across the NDA group, ensuring the foundation remains compliant, explainable, and adaptive to future regulatory standards.
“Good data is not an output — it’s an asset.
The SATNAV framework proved that trust in AI begins with trust in data.”
— Surtori Principal Consultant, Data & AI Strategy