- Data Sovereignty
Building the Global Infrastructure for Professional Judgment in an Agentic World
The vision of a truly data-driven system is no longer a distant goal; it is the immediate frontier. In 2026, we are witnessing a fundamental shift in how organisations understand data and expertise. This isn't just a story for schools; it is a global movement toward Sovereign Intelligence that spans every sector.
We see this shift clearly in two recent landmarks in the Education sector:
- The vision for a data-driven school system, prioritising interoperability and the death of proprietary silos (https://www.openeducationai.org/).
- The new era of Welsh educational data, which replaces "black-box" scores with professional agency and holistic evidence.
A Universal Challenge: Beyond the Classroom
The themes we see in education such as data fragmentation, lack of trust in "automated" grades, and the need for authentic evidence, are identical to those in recruitment, awarding, training and compliance.
Across every sector, the old model of a "Final Mile" service is becoming a bottleneck. In an agentic world, where autonomous AI agents monitor performance and trigger interventions, organisations cannot afford to have their "judgment data" trapped in a vendor's silo. Providers need the invisible, high-fidelity infrastructure that feeds validated data into the customer’s own private cloud.
The key Questions for Future-Ready technology providers
Whether you are a MAT, a global bank, or a national ministry, you must ask your software providers these key questions to ensure you aren't being locked into a legacy silo.
| The Question | RM Compare (Infrastructure) | Typical SaaS (Service) |
|---|---|---|
| API First: Can I access every data point via a modern API? | ✅Roadmap: We allow full data export now; external APIs are currently in development to allow seamless machine-to-machine integration. | ❌No. Primarily a closed web-portal experience. |
| Sovereignty: Can I run my own AI agents against the data in my private cloud? | ✅Yes: Our data exports are structured to be piped directly into your own AI environment. | ❌No: You use their AI and their benchmarks. |
| Data Ownership: Do you use my candidates' work to train your proprietary models? | ✅No: We are a data processor; you are the controller. Your data is never used to train our internal models. | ❌Yes: User data is often aggregated. |
| Interoperability: Does your data model use a graph-based approach? | ✅Roadmap: Our core technical stack is built for this; the implementation of a full data graph is a part of our active roadmap. | ❌No. Relies on traditional relational database structures. |
| The "Final Mile": Can I apply my own contextual logic to the raw data? | ✅Yes: We provide the "Rank and the Ruler." You decide what those measurements mean in your specific context. | ❌No: They provide the context |
| Edge Capability: Are you researching on-device AI to protect privacy? | ✅Yes (Roadmap): We are actively researching in-browser AI to ensure sensitive data never leaves the user's device. | ❌No: Heavy reliance on centralised, server-side processing. |
| Multi-Modal Capture: Can your system handle video, audio, and physical work? | ✅Yes: Our Companion App is designed specifically for these "un-markable" forms of evidence. | ❌No: Optimized for easy to assess materials |
| Statistical Transparency: Do you provide raw reliability metrics (Standard Error)? | ✅Yes: We offer full transparency on the health of your assessment; no "black box" scores. | ❌No. Focus is on the final score rather than raw technical audit |
| Privacy by Design: Is your privacy policy built for "Data Controllers"? | ✅Yes: Our Privacy Approach is built for organisations managing their own risk. | ❌No. Built for local-level GDPR compliance. |
| Agentic Readiness: Is your data structured for AI agents to act upon today? | ✅Yes: Our validated, structured data exports are "machine-ready" for consumption by your internal agents. | ❌No. Data is presented for human consumption on THEIR platform. |
| Governance & Tenancy: Can I manage multiple "tenants" or branches with centralized oversight? | ✅Yes: Through our License Centres and Connectors, we allow complex organizations to manage data flow across multiple sub-entities. | ❌No. Optimized for individual sites |
| Security & Compliance: Do you meet international standards for data security and infrastructure? | ✅Yes: We are backed by ISO 27001 certification and Cyber Essentials Plus, ensuring your data is protected by gold-standard security. | ❌No. Generally ok for local level, small scale compliance |
| Proven Experience: Is your technology backed by deep, trusted experience in high-stakes assessment? | ✅Yes: RM Compare is part of RM Assessment, a global leader with decades of experience delivering secure, high-stakes assessments for governments, ministries, and awarding bodies worldwide. | ❌No. Limited experience locally, regionally and internationally/ |
Our Roadmap: The Architecture of Expertise
To understand how RM Compare is building this infrastructure, it helps to look at our three-stage roadmap journey. This is not just a plan for new features; it is a strategy to transform raw human judgment into a high-utility data stream.
1. The Capture Layer (Now)
The first challenge in any assessment is getting high-quality evidence into the system without friction. We are solving this through our mobile-first ecosystem. Our Companion App removes the "cables and uploads" bottleneck, allowing you to digitise authentic, multi-modal evidence (for example speech, physical portfolios, or performances) at the moment they happen.
The Goal: To "capture the un-capturable."
2. Intelligent Insight (Next)
Once the data is captured, we move beyond the scorecard. By leveraging our Adaptive Comparative Judgment (ACJ) engine, we convert thousands of individual expert decisions into highly reliable rank orders. We are building the tools that allow this data to feed directly into your private AI agents, providing insights that flag marking bias and identify learning gaps automatically.
The Goal: To provide "machine-ready" data that tells you what to do next.
3. Professional Compass (Later)
The final stage of our journey is the Professional Compass. This is where RM Compare moves from an assessment tool to a governance and development ecosystem. Through our License Centres and Data-Sharing tools, large organisations can build a shared, internal "Compass" of quality, ensuring a "Gold Standard" remains consistent across 50 schools or 10 global offices.
The Goal: To provide a universal sense of direction for professional judgment.
How RM Compare and RM Assessment are Responding
We are well-placed to pivot to this change because our architecture was built for it. Our roadmap is focused on ensuring we provide the most reliable measurement engine in the world.
- Sovereign AI Support: We assume the AI will be owned by the customer.
- Enterprise Governance: We use tools like Connectors and License Centres to allow large-scale organisations to scale assessment without losing oversight.
- The Graph: We are moving toward a full graph-based data model to allow for the complex, multi-dimensional queries that modern AI agents require.
RM Compare is a specialised engine within a broader response from RM Assessment. While we focus on the infrastructure of judgment, the RM Ava platform is building the "Operating System" for the entire digital assessment lifecycle. While we handle the "un-markable" evidence through Comparative Judgment, RM Assessment ensures that high-stakes awarding bodies have the secure, trusted, and legal framework required to deliver formal qualifications at scale.
Conclusion: You Drive the Car
The "Agentic World" demands a new kind of partner. One that prioritises Sovereignty, Interoperability, and Technical Excellence, backed by Deep Institutional Trust. We are building the "Intel Inside" for professional judgment.
We provide the engine; you drive the car. See our Roadmap for the journey ahead.