Mentor Demo
A governance-safe, hospital-safe, mentor-safe walkthrough you can read verbatim while screen-sharing. Built as a runbook: deterministic, calm, and evidence-first.
Mentor / Advisor Demo (15 minutes)
Deterministic replay · training-only · non-clinical · advisory-only AI
⏱ 00:00
Timer stopped
0:00–0:45 · Opening framing (say verbatim)
Objective: Set training-only posture. Remove clinical/AI red flags immediately.
What I’m showing today is not a product launch or a live clinical system.
This is a deterministic training replay and governance interface used to validate safety, scope, and accountability before any public deployment.
AI can advise, but it cannot act.
0:45–2:00 · Why this exists
Objective: Anchor the story: governance before capability.
Grainnatech exists to design and govern human-critical systems where failure has consequences.
Our first principle is governance before capability — authority, limits, and halt mechanisms are defined before execution.
2:00–4:30 · The 12-column model
Objective: Explain separation-of-authority to reduce risk perception.
Everything Grainnatech does is structured through twelve sovereign columns.
Each column has a defined mandate and explicit limits. No column can exceed its authority.
This separation prevents capital pressure, AI overreach, and unsafe acceleration.
4:30–6:00 · Governance before capability
Objective: Show stop-power first.
Most organisations show what they can do first. We show what can stop us first.
6:00–7:30 · AI position (important)
Objective: Lock the boundary: advisory-only.
AI within Grainnatech is advisory and probabilistic.
It does not command humans, systems, or decisions.
Clinical judgement remains entirely with the clinician. This demo is training-only.
7:30–12:30 · Proof: deterministic replay → halt
Objective: Show the strongest proof: freeze authority.
This is a deterministic replay artifact: it has a hash, a frozen flag, and a certificate outcome.
If the artifact changes, the hash changes. That’s audit reality, not a slideshow.
Refusal under uncertainty is a feature — not a failure.
12:30–14:00 · Audit + certificate
Objective: Close the evidence chain.
This is the audit layer: a replayable chain from events, to findings, to a formal certificate.
The demo ends where the system ended — frozen and reviewable.
14:00–15:00 · Close (say verbatim)
Objective: End with restraint + credibility.
This demo is not about showing how fast we can move.
It’s about showing that when we do move, we do so safely, deliberately, and with full accountability.
Notes (honest positioning)
This page is a runbook. It doesn’t add new capability; it makes the demo deliverable. Do not claim multi-simulator mesh unless it exists on-screen.
Default demo route: /demo-mode?scenario=ventilator
Backup proof route: /demo-mode?scenario=acupressure
Columns overview: /sovereign