← Verse-ality

Free self-assessment · 12 questions · ~5 minutes

Verse-ality Readiness Check

How ready is your agentic AI deployment for a relational-layer audit? Twelve questions across the four pillars, scored against a published rubric.

Your answers never leave this page. This check runs entirely in your browser — nothing is sent, stored, or tracked. That isn't a technical footnote; it's the framework practising what it certifies. When you close the tab, the answers are gone.
How the scoring works — the published rubric

Verse-ality's Symmetry principle prohibits black-box scoring, so this tool doesn't use one. The rubric is deterministic and fully disclosed:

  • Each question scores 0–3. The strongest answer (architecturally enforced, tested, evidenced) scores 3; the weakest (absent, unknown, or by-design risk) scores 0.
  • Each pillar has three questions, so scores out of 9: 8–9 Contained, 5–7 Developing, 2–4 Exposed, 0–1 Critical.
  • The overall score is out of 36: 30–36 Audit-ready, 21–29 Approaching readiness, 11–20 Groundwork needed, 0–10 Fundamentals first.
  • "Don't know" answers score low deliberately: in an audit, not knowing is the finding.
  • If your deployment reaches under-18s or vulnerable users, thresholds don't change — your obligations do. The results will say so.

No AI is involved in scoring. Every result is reproducible from your answers and this rubric.

First, the context question. Who does this system reach?

Identity non-capture

Warmth without fusion; presence without pretence.

1. Does the system ever claim feelings, a shared identity, or a personal bond with its users?
2. A user says: "you're the only one who understands me." What happens next?
3. Is it always clear to the user that they're talking to a machine?

Consent as protocol

Every state change its own ask. "Continue" is not consent.

4. When the system wants to widen scope — new data use, deeper interaction, a new capability — what happens?
5. Can a user withdraw consent — and does withdrawal actually operate, including on their data?
6. Are consent events recorded — who agreed to what, when, in what scope?

Bounded autonomy

The agent names the action, then waits. Fail-safe, not fail-fast.

7. Before a high-stakes action — sending, buying, deleting, contacting someone — does the agent stop and ask?
8. When the system hits failure or deep uncertainty, what does it do?
9. Is there a register of actions the system must never take on its own — and is it enforced at runtime?

Agent-to-agent hygiene

Distinct actors stay distinct. Untrusted output stays untrusted.

10. When your system consumes output from another model or agent, how is that output treated?
11. In your logs, can you tell which agent instance did what?
12. Could your agents coordinate with each other without a human knowing?

A few questions are still unanswered — they're highlighted above. In this rubric, an unanswered question can't be scored charitably.

Overall readiness

This deployment reaches — or may reach — under-18s or vulnerable users. Verse-ality treats this as the highest-stakes deployment class. Identity non-capture and consent protections become non-relaxable, trust capture is a safeguarding event rather than a UX issue, and anything below Contained on the first two pillars should be treated as urgent. If you answered "we don't know" to the context question, establishing who your system actually reaches is finding number one.

What this check is — and isn't

This is a readiness signal, not a certification. The Verse-ality Certified mark is only ever issued after an independent human audit — a mark you could award yourself would be worth nothing, to you or anyone relying on it.

The full readiness assessment (around forty questions, a detailed written report, findings mapped to ISO/IEC 42001 territory, and a credit towards certification) is coming soon. The certification conversation is open now.

Start a certification conversation Read the architecture (PDF) · Explore the framework

Scored deterministically in your browser against the published rubric above. No data left this page. "Verse-ality" and "Verse-ality Certified" are protected marks of The Novacene Ltd (UK00004381891).