Sovrenticv1
Request early access
Research & Standards

The evidence base behind the Sovrentic assurance scheme

Sovrentic is being developed as an inspectable assurance scheme for the agentic era, against a visible body of standards, academic research and industry practice covering assurance, identity, delegated authority, governance, evidence and digital sovereignty.

These sources inform our method. They do not constitute endorsement, conformance or proof of compliance.

01 / Sovrentic Papers

From research to verifiable assurance.

Sovrentic translates research and standards into a proposed assurance framework for agents, tools and skills. Our SOV paper explains how identity, declared authority, assessment evidence and current status can support decisions across technology and GRC.

Technical working paper · v0.2.1

Sovereign Operational Verification: A Change-Aware Assurance Layer for Agentic Resources

Organisation
Sovrentic
Version
0.2.1
Date
10 September 2026
Review
Not peer reviewed
Language
English

The SOV paper proposes a change-aware assurance lifecycle for agentic resources. It connects resource identity, delegated authority and scoped evidence, then asks what remains supported when a model, tool, permission or operating context changes. The paper defines a continuity procedure, documents prototype limitations and sets out an evaluation plan.

A technical proposal with an unexecuted comparative evaluation plan. It does not establish an accredited or production assurance scheme.

Read paper ↗Download PDFPaper details →
02 / Method

How we use these sources

  1. 01

    Define assurance dimensions

    What should be assessed, and within which scope?

  2. 02

    Model identity and authority

    Who is the agent, who controls it and what was delegated?

  3. 03

    Preserve verifiable evidence

    What was tested, by whom, when and under which conditions?

  4. 04

    Monitor change and sovereignty

    What remains valid when systems, dependencies or jurisdictions change?

03 / Reference register

Core references

A working register of materials reviewed for relevance. Inclusion records influence, not adoption, endorsement or conformity, and no conformance to any listed standard is claimed.

Standards we align to

Published standards and regulation whose structure informs the SOV schema. Alignment status: informing our method, not assessed for conformance.

Official standard2023

ISO/IEC 42001

Frames how an organisation can establish and improve an AI management system.

Status · Published
View source ↗
Official standard2023

ISO/IEC 23894

Provides risk-management guidance for organisations developing or using AI.

Status · Published
View source ↗
EU regulation2024

EU AI Act

Sets risk-based obligations and governance expectations for AI in the European Union.

Status · In force
View source ↗
Technical standard2022

W3C DID Core

Defines decentralised identifiers and verification methods for resolvable identity records.

Status · Recommendation
View source ↗
Technical standard2025

W3C Verifiable Credentials Data Model

Defines an interoperable data model for expressing and verifying digital credentials.

Status · Version 2.0 recommendation
View source ↗

Work we reference

Working drafts, academic research and industry frameworks we track. Referencing this work implies no endorsement by its authors and no conformance by Sovrentic.

Public framework2023

NIST AI RMF

Offers a practical vocabulary for governing, mapping, measuring and managing AI risk.

Status · Version 1.0
View source ↗
Working draft2026

IETF Agent Identity Protocol

Explores a protocol for identifying agents and binding their identity to accountable principals.

Status · Internet-Draft · non-normative
View source ↗
Working draft2026

IETF AI Agent Authentication and Authorization

Explores authentication and delegated authorisation patterns for AI agent interactions.

Status · Internet-Draft · non-normative
View source ↗
Industry framework2017–present

SPIFFE and SPIRE

Provides workload identity patterns relevant to authenticating software agents and services.

Status · Open-source project
View source ↗
Industry framework2022

CertAI

Offers an industry example of structured validation for trustworthy AI systems.

Status · Commercial validation service
View source ↗
Academic research2024

Implementing a Certification Scheme for AI Trustworthiness

Examines how trustworthy-AI criteria can be translated into a structured assessment scheme.

Status · Peer-reviewed paper
View source ↗
Academic research2026

TrustBench

Studies real-time trust verification before an autonomous agent executes an action.

Status · Research preprint
View source ↗
Industry framework2026

Cloud Security Alliance Agentic Trust Framework

Applies zero-trust governance concepts to the identity and operation of AI agents.

Status · Industry publication
View source ↗
Academic research2019

Model Cards

Introduces structured reporting of model purpose, evaluation conditions and limitations.

Status · Peer-reviewed paper
View source ↗
04 / Governance

Our editorial rule

Sovrentic distinguishes between official standards, regulations, technical drafts, academic research, industry frameworks and market commentary. Every reference is versioned, scoped and reviewed before it is used to inform a product control or assurance profile.

A reference can inform our method without proving that an agent is safe, compliant or trustworthy in every context.

Last reviewed: 9 September 2026

Continue the conversation

Examine the method with us.