Sagolik Institutional · Proposed product

AbrahamMoses Decision Twin

Evidence Today. Wisdom Tomorrow. A structured intelligence system for decisions that must withstand scrutiny. It models a decision, not a person.

Evidence informs. Humans decide. The system preserves the reasoning.

Implementation status

What exists today

The Decision Twin is a proposed product. Today it exists only as a public demonstration on one fictional private credit case: the Decision Room and its synthetic Decision Passports. There are no private workspaces, accounts, customers or customer data. No bank, fund or institution has reviewed, validated or used it, and it holds no certification. The only thing for sale is a paid pilot reservation: the 90-day pilot starts once private workspaces are available.

A Decision Twin is a digital representation of one decision: its evidence, assumptions, alternatives, risks and approvals, kept as a versioned record of why it was justified, what could invalidate it and who approves it. It models a decision, not a person.

Architecture

Six layers, and which are demonstrated now

The six layers of the Decision Twin and their status
LayerWhat it doesStatusToday
1. Data ingestionAuthorized PDFs, contracts, financial statements, papers, APIs and datasets.PlannedNo upload or connector exists. The Decision Room uses 10 fixed synthetic documents written into the code.
2. Evidence engineAtomic claims, passages, citation checks and provenance.Partly demonstratedOperational for the public AbrahamMoses research record (claims, sources, revisions). In the Decision Room each synthetic document has a stable ID and SHA-256 hash.
3. Decision Twin graphLinks the decision to claims, assumptions, calculations, counterarguments and alternatives.Demonstrated (synthetic)Dependency graph for one synthetic case; changing an input recalculates only its dependents.
4. Analysis and simulationDeterministic calculations plus controlled AI: scenarios, challenges and contradictions.Partly demonstratedDeterministic calculations, scenarios, rule-based Adversarial Chamber and Contradiction Map are demonstrated. No AI is used in the Decision Room.
5. Governance and approvalRoutes to authorized humans, keeps version history, records approvals and objections.PlannedNo accounts, roles or approvals exist. The demonstration shows where human judgment is needed and never records an approval.
6. Decision PassportReproducible record of sources, assumptions, calculations, review history and integrity metadata.Demonstrated (synthetic)Stateless, synthetic passports with a SHA-256 hash and JSON, Markdown and print exports. Reviewer: none (simulated decision).

AI interpretations, when they are added, will stay distinguishable from verified information and from deterministic calculations. The demonstration uses no AI.

Configuration

Per decision, and per workspace

Each decision carries its own configuration: the objective, the evidence standard, which assumptions are fixed, adjustable or disputed, the approval policy (propose, review, reject, approve), escalation rules for missing or contradictory evidence and monitoring triggers. It is saved with every version so the decision can be reproduced. The Decision Room shows it for the synthetic case, and every Decision Passport contains a snapshot.

Default workspace settings (planned for private workspaces; read-only)

No private workspace exists yet; these defaults are not in effect anywhere.

Workspace visibility
Private
Access control
Role-based, least privilege
Tenant separation
Strict
AI providers
Approved providers only
Data region
Customer-selected supported region
Source permissions
Per document and per user
Automated decisions
Disabled
Consequential actions
Human approval required
High-risk decisions
Two-person review
Audit log
Tamper-evident, on
Retention
Customer-configurable
Training on customer content
No, unless explicitly authorized
Evidence monitoring
Opt-in, scoped to named sources
Export
Role-restricted
External integrations
Off until authorized

Two environments

AbrahamMoses Scholarly and Sagolik Institutional

AbrahamMoses Scholarly is the public research record on this site. Sagolik Institutional, the proposed home of the Decision Twin, would hold confidential data under governance, simulations, controlled integrations and contractual service levels. They would share technology but keep access, data and editorial responsibility separate; religious research can be sensitive, so public and institutional workspaces stay apart. More on the difference.

Pricing

Plans

Prices exclude VAT; the seller is not VAT-registered. The pilot is a paid reservation that starts once private workspaces are available.

Public institutional demonstration ready — private client onboarding not authorized.

Pilot

€9,500 one-off reservation, 90-day pilot

  • One private workspace
  • Up to 10 users
  • Onboarding

Reserve a pilot

Reservation: the 90-day pilot starts only when private institutional workspaces exist; the start date is agreed with you in writing.

By reserving you accept the Terms of Sale (not reviewed by a lawyer). Full refund on request if the pilot has not started within 90 days of payment. Questions first? Ask us by email.

Institutional

€2,900 per month, billed annually

  • Unlimited Decision Twins
  • 25 users
  • Exports and API access

Talk to us

Enterprise

Custom by agreement

  • Choice of data region
  • Single sign-on (SSO)
  • Service-level agreement
  • Contract terms for banks, including DORA requirements

Talk to us

Seller: Sagolik, Halmstad, Sweden. Contact: markus@sagolik.com. Prices exclude VAT; the seller is not VAT-registered. A pilot is a paid reservation that starts once private workspaces are available; payment is taken by Stripe, which issues the receipt. Decision support only: the customer keeps responsibility for its decisions, and a Decision Passport is not a binding decision, audit opinion, regulatory approval or certification. No DORA assessment or regulatory approval is claimed; a written agreement is required before any client data is used.

Open the Decision Room (synthetic demonstration)