The Sovereign Exit: Why Portability is the Ultimate Trust Test

By Mark House

In our series on digital sovereignty, we have explored how "Harvesting Models" use dark patterns to bypass consent and how "AI Wrappers" quietly extract value from your institutional judgment data. But there is one final, definitive test of a vendor’s intent: What happens when you want to leave?

If a platform is truly sovereign, your departure should be as simple as a data export. If it is a harvesting model, your exit will feel like a ransom negotiation.

The "Data Prison" vs. The "Sovereign Vault"

Many assessment tools create a "closed loop" where the real value - the benchmarks, the longitudinal insights, and the calibrated "Rulers" - exists only within the vendor's portal. This is Vendor Lock-in masquerading as a service.

  • In a Harvesting Model: You can often export a basic PDF of results, but the underlying "intelligence" - the machine-ready data that fuels AI agents - remains proprietary to the vendor. You have rented their engine, but you cannot take the "map" with you.
  • In a Sovereign Model: Your data is structured to be "machine-ready" for your ecosystem. Because the platform is built on infrastructure rather than extraction, you own the raw material and the derived insights.

The 3 Requirements of a Sovereign Exit

To ensure you aren't being "locked in" to a dependency trap, any "AI-enabled" tool must meet three portability standards:

1. Full Semantic Portability

It isn't enough to export a spreadsheet of scores. A sovereign tool allows you to export the entire judgment graph . That is the relationships between scripts, the comparative metadata, and the calibrated standards.

  • The Test: Can we move our "Institutional Ruler" into our own private AI environment without losing its fidelity?

2. Machine-Ready Streams

In an agentic world, your data must be ready for your agents to act upon.

  • The Test: Does the vendor provide an API or export format that allows our own internal models to learn from our historical judgments?

3. No "Intellectual Property" Tail

When you leave a sovereign partner, they should retain zero rights to your data.

  • The Test: Does the contract explicitly state that all "worldwide licenses" for benchmarking or research are revoked upon the termination of the service?

Why Portability is the Ultimate Trust Test

The willingness to make data portable is a character reveal. A vendor that builds "walls" to keep you in is a vendor that views your data as their asset. A partner that builds Sovereign Infrastructure knows that their value lies in the quality of their engine, not the capture of your fuel.

At RM Compare, we don't just give you the keys to the car; we ensure you own the engine and can drive it wherever you choose. Our goal is to build a comparative judgment engine that converts human expertise into high-fidelity data streams that stay in your ecosystem.

The Sovereign Standard means having the power to say "yes" to collaboration, and the absolute freedom to say "goodbye."

The Sovereign Exit Checklist

1. Data & Semantic Portability

  • Full Graph Export: Can you export the complete judgment graph, including all relationships between scripts and metadata, or only a flat results table?
  • Machine-Ready Formats: Is data provided in formats (like JSON or structured CSV) that allow it to be immediately ingested by your own internal AI agents or "graph" environments?
  • Derived Data Ownership: Do you own the "Institutional Ruler" (the calibrated standards) created during the process, or does that intelligence stay locked in the vendor’s portal?

2. Intellectual Property & Licensing

  • License Revocation: Does the contract explicitly state that any license granted to the vendor (even for "anonymised research") is fully revoked upon account termination?
  • Zero-Residual Training: Can the vendor guarantee that your students' work has not been used to train "global" or "collective" models that they will continue to profit from after you leave?
  • Copyright Affirmation: Does the agreement recognise that pupils usually own the copyright to their work and that you, the institution, are not authorizing its permanent commercialization?

3. Operational & Jurisdictional Freedom

  • Migration Support: Is there a documented process for bulk exporting all raw and derived data without incurring "ransom" fees or administrative hurdles?
  • Fixed Residency Audit: Can you verify that no part of your "Sovereign Exit" involves routing data through overseas sub-processors (like US-based AI stacks) during the export process?
  • Platform Independence: Can the core assessment logic (the comparative engine) be run on your own stack or tenancy, or are you permanently dependent on the vendor's cloud?

The Verdict: Are You "Locked In"?

If you answer "No" to more than two of these questions, you are likely in a Harvesting Model. You are paying for a service that treats your professional judgments as its own proprietary fuel.

RM Compare is built to pass this checklist. We provide the engine, but we ensure you always own the "Institutional Ruler" and have the absolute freedom to take your intelligence elsewhere.