From waiting to moving

From scattered evidence to a typed decision a human can review.

Before, a prepared contract exception waits for the CRO, CISO and General Counsel to reconstruct context. After, the choice, evidence gap, policy gate and named reviewer arrive together while human authority remains intact.

Before

Someone hunts for the policy, the precedent and the prior determination.
The rationale behind the last decision is reconstructed from memory.
The decision then waits again for an approval slot.

Intended after state

Prior reasoning is structured once and reused under governance.
Evidence and its provenance travel with the answer.
Review and escalation sit inside the workflow, not after it.

The business consequence is fewer repeat reconstructions and decisions organized around evidence rather than calendars.

The decision is structured before it is released.

This sample contract data-use example separates an enum choice, rubric score, binary probability, deterministic authorization and human release. Probability and confidence do not grant policy permission.

Static modeled schema

Commercial contract data-use exception

A CRO, CISO and General Counsel review a requested customer-data use before signature.

Decision question: May this contract proceed with the requested customer-data use?
enum_choice: HOLD
rubric_score: Not evaluated
risk_band: HIGH
evidence_complete: false
binary_probability: null
confidence: null
calibration_status: Not evaluated in this sample walkthrough
deterministic_authorization: Missing authorization
human_release: Named General Counsel. Await human review

Demo walkthrough. This is not model execution, a benchmark, a calibrated output or a customer result. Probability and confidence are not policy permission. An uncertain model score cannot override deterministic authorization or mandatory human review.

How the change happens: Ingest, Twin, Operate.

Only after the change in operating state does the product matter.

Ingest

Authorized enterprise evidence enters under the permissions and controls of the selected deployment architecture, and is reconstructed into structured decision episodes.

Twin

Episodes become a governed, role-specific judgment model: principles, thresholds, exceptions, precedent, authority boundaries and escalation rules, connected by the Role Wisdom Graph.

Operate

The twin becomes available through One-Press Huddle and other governed interfaces. Each response carries evidence, provenance, permission and policy context, and resolves to Recommend, Escalate or Safe Decline.

Ingest

Evidence is designed to enter under the permissions and controls of the deployment pattern selected during pilot scoping.

Sources

Prior decisions and their outcomes
Policies, standards and procedures
Committee minutes and determinations
Approved communications and memos
Systems of record and structured data

Decision Episodes

The question that was asked
The evidence that was considered
The constraints that applied
The authority that decided
The outcome and the rationale

Controls at intake

Source-level entitlements preserved
Classification retained with the record
Provenance attached to every episode
Perimeter set by the deployment pattern chosen during pilot scoping

Twin

Episodes become a governed, role-specific judgment model connected by the Role Wisdom Graph.

What the twin encodes

Principles the role applies consistently
Thresholds that change the answer
Exceptions and the conditions that justified them
Precedent and how it was weighted
Authority boundaries and delegated limits
Escalation rules and named validators

What the twin is not

Not a clone of a specific person
Not a digital replica or voice copy
Not an autonomous employee
Not a document search interface
Not a decision maker: a human owns the decision

Operate

The twin becomes usable through One-Press Huddle and other governed interfaces.

Recommend

Evidence is sufficient, policy checks pass and the answer stays inside the role's authority.

Escalate

The decision exceeds the role's authority or needs a named human validator before it moves.

Safe Decline

Evidence or permission is insufficient. The platform says so and routes the decision to a human.

Every response carries evidence, provenance, permission and policy context. Where evidence or permission is insufficient, the design calls for a Safe Decline rather than an unsupported answer.

Why judgment is not just chat

TypeSafe introduced Jev in September 2026 as a System One model for typed decisions. That is an external example of a shift from generated prose toward structured decision outputs. WisdomTwin.ai is pursuing the enterprise workflow around that shape: role judgment, authorized memory, deterministic controls and human release inside an approved deployment boundary. We do not claim a TypeSafe partnership or a Jev integration.

TypeSafe publicly documents a hosted API and US hosting. We did not locate a public on-premises Jev offering in the reviewed launch materials. That is not a claim about all possible enterprise arrangements. Its privacy policy says inputs are not used to train or fine-tune its AI.

Primary sources

Start with the decision that waits most often.

In one working session, we map where a recurring decision stalls today, which role holds the judgment, what evidence and approvals govern it, and what the accountable after state should be.

Request an Enterprise Pilot See the Before and After