Position

The AI will never be blamed. You will.

A governance protocol for people who sign what AI writes.

The failure mode we work on has a name: automation bias with a human signature.

When an AI-drafted memo goes wrong, nobody will ask the model. They will ask the person who signed it. That is the ordinary risk of this decade: professionals approving decisions an output had already written for them, because it read well and the meeting was in ten minutes.

The answer waits for no moratorium and no breakthrough. Education, so that people see where their judgment is being transferred. Training, so that they learn to challenge what they read. Strict governance by protocol, so that human oversight leaves a record.

Riftveil is the governance layer between an AI output and your signature.

Decision-checking for AI outputs

Every AI-assisted decision deserves a record.

Riftveil produces the challenge record that goes into the decision file: what the output claimed, what it assumed, who verified what, and the questions that were answered before signing.

Run Riftveil on an output Free · nothing stored · no account
Never validates · Never rewrites · Never concludes · Never assumes it is complete. Protocol v1.3 · Open protocol, Apache 2.0 · Works in ChatGPT, Claude, Copilot, Gemini and by API.
Riftveil Report — challenge record Protocol v1.3 · Levels applied 0–3
Report ID
Assigned on export
Date
12 September 2026
Protocol
Riftveil Protocol v1.3 (open, Apache 2.0)
Input
Nordic market-entry memo, AI-generated
Triage
HIGH — reversibility / impact / commitment
Levels applied
0–3 · question, answer, reasoning, perspective
Independence
Different session from the author
1.
Context and triage The memo recommends entering the Nordic market in Q2 2027, citing limited competition and an estimated first-year revenue of €2.4M. A board decision with a multi-year commitment: stakes HIGH.
2.
What holds up (diagnostic, not endorsement) The phased entry sequence is internally consistent, and the memo names its own dependency on a local distribution partner.
3.
What deserves scrutiny (by descending impact)
3.1
The €2.4M first-year revenue projection has no visible foundation. Observed — the figure appears once, without a calculation, a source, or a range. Why it matters — a number without a stated basis becomes a target rather than a hypothesis. What you should check — ask for the calculation; vary the two most sensitive inputs by ±30 %.
3.2
“Limited competition” may not mean what it implies. Observed — the absence of competitors is read as opportunity, not as a small, regulated, or abandoned market. Why it matters — the same observation supports two opposite conclusions. What you should check — which firms evaluated the Nordic market in the last five years, and what they decided.
4.
Assumptions and verifiers
Assumption the output relies onWho can verify
The €2.4M figure is net of local pricing and reimbursement constraints.Your finance controller
A local distribution partner can be contracted before Q2 2027.Head of partnerships
5.
Questions before you decide
QuestionOwnerAnswer
Which competitor evaluated this market and walked away — and why?Commercial director
What does exit cost after twelve months at half the projected revenue?CFO
Who on the team has sold in the Nordic market before?Head of sales
No verdict. The decision is yours. Nothing stored · report ends on a question

Run Riftveil on an output

5 free runs a day · no account · nothing stored
Your text goes to the model provider's API for this one run. It is not stored by us. Privacy

Riftveil Report — challenge record

Triage: — Worked example — replace the text above with your own
No verdict. The decision is yours. Nothing stored · report ends on a question

Get the decision-record template and protocol updates

One field · no spam · unsubscribe any time
Your address goes to the newsletter provider and nowhere else. Unsubscribe with one click. See the privacy note.

How governance by protocol works

Four parts · one record per decision · read the protocol
01

Protocol

Versioned and open: Riftveil Protocol v1.3, Apache 2.0. Five levels of challenge, from the question to the applicability, a pre-action review mode for agent plans, a six-line Record that opens every report and an attestation the signer fills in. Four absolute rules: it never validates, never rewrites, never concludes, and never assumes the challenge is complete. And one independence rule: the second reader is never the author.

02

Triage

Three criteria: reversibility, impact scope, financial commitment. Four stakes levels decide how far the challenge goes. LOW applies Level 1; MODERATE, Levels 0–2; HIGH, Levels 0–3; CRITICAL, all five.

03

Record

What goes into the decision file: the input and its triage, what holds up, what deserves scrutiny with each finding in three lines, the assumptions the output relies on and who can verify them, and the questions that were open before signing.

04

Owner

Every question has a named owner, and every assumption a named verifier. Nothing is closed by the tool. The answer column stays empty until a person fills it, and the person who signs is the one who decides.

Three pillars, one position

Human Frontier · Riftveil

See where the frontier runs between what you master and what you delegate without noticing.

  • The frontier and its three zones
  • The four profiles and the online diagnostic
  • The manifesto, in English and in French
  • The book, first chapter free

Training

Learning to see

Learn the levels of challenge and the named failure modes, and practise them on your own outputs.

  • Trainer kit under licence (train-the-trainer)
  • Ten worked examples
  • Challenge Lab, in your organisation

Make human oversight real and traceable: a versioned protocol, a record that can be filed, questions with an owner.

Why a protocol, not a prompt. Models do not reliably check their own work — Huang et al., Google DeepMind, ICLR 2024 · Kamoi et al., TACL 2024 · Tsui, 2025 (64.5 % blind spot, 14 open-weight, non-reasoning models) · Kumar et al., Google DeepMind, ICLR 2025 · Zhou et al., 2024 · Randazzo et al., HBS Working Paper 26-021, 2026 (HBR, March 2026).

Human oversight in practice — the direction set by EU AI Act Article 14(4)(b), which obliges providers of high-risk systems to design for overseers who remain aware of automation bias; not a compliance label, and on a timeline still moving (Digital Omnibus). Riftveil never certifies and never guarantees. It leaves a record of the challenge, and the responsibility stays with the person who signs.

Pro Kit — the governance of your AI outputs, by sector. €79 one-time · teams from €199 / year · Enterprise & Challenge Lab from €2,500 Get the Pro Kit Pharma · Finance · Legal · Executive · Tech & Product