AN OPEN FRAMEWORK BY NIHARIKA SRIVASTAV AND SANJAY SAXENA

VERDICT — A Framework for
Governing Agentic AI.

Seven pillars for safe, compliant, enterprise-ready AI agents. Open. Free to use. Built to be defended to a board.

The Seven Pillars

V
Validation
Testing, red-teaming, conformity before deployment.
E
Evidence
Audit trails, sign-offs, lineage, provenance.
R
Runtime Control
Guardrails, kill-switches, human-in-the-loop, containment.
D
Decisions
Reasoning firewalls, decision logging, explainability.
I
Identity
Non-human identity, signed agent cards, scoped authority, zero-trust.
C
Cost & Compliance
Token budgets, execution caps, policy-as-code, regulatory mapping.
T
Transparency
Model cards, fairness metrics, disclosure.
The mnemonic spells VERDICT. The build order starts with Identity - you cannot govern what you cannot name.
[REF-01]

What VERDICT Is

VERDICT is an open, seven-pillar framework for governing AI agents and agentic systems through their full lifecycle - from inventory to runtime control to self-governance. It is deliberately incomplete: a living standard meant to be extended by the people who use it.

Open and Free

VERDICT is released under Creative Commons Attribution 4.0 (CC BY 4.0). Free to use, cite, and build on, with attribution. Certification, training, and tooling are separate.

VERDICT framework. Authored by Niharika Srivastav and Sanjay Saxena.
First published June 7, 2026. Licensed CC BY 4.0.

The 74th Concept

VERDICT defines 73 concepts. The framework is open by design: the 74th is yours.

Contribute the 74th concept - and get notified when the book ships.

No spam. Launch news and framework updates only.

[REF: THE BOOK]

Governing Agentic AI

The book that builds VERDICT across one year inside a Fortune 500 bank.
Coming soon.

VERDICT FRAMEWORK
AUTHORS: NIHARIKA SRIVASTAV AND SANJAY SAXENA
© 2026 VERDICT. LICENSED UNDER CC BY 4.0.
CONTACT: [email protected]

[SYS_REF: CANONICAL_FOOTER // UPDATED: 2026-06-07]