CognitionHive | CRI Research →
Wealth Intelligence · In Preview
Nexum Prism

Portfolio evidence you can
inspect before you act.

Nexum Prism is the planned self-directed investor experience for the Nexum architecture. It is designed to make forecasts, constraints, allocations, and trades visible as distinct evidence rather than presenting a single black-box score.

For institutions: Nexum Fiducia →
A CognitionHive product. Built from the Nexum engineering architecture and Claim Reliability Index framework. Nexum Prism is pending USPTO registration in Classes 36 and 42.

A product experience is planned; the evidence engine is implemented

Implemented Engine

A private engine creates deterministic forecast, constrained-allocation, trade, recommendation, replay, and four-dimension CRI artifacts over synthetic data.

Planned Investor Experience

The Prism account experience, real portfolio connections, market-data pipeline, authentication, monitoring, and operational deployment remain to be built.

Human Decision

The architecture is designed to expose evidence for review. It does not establish suitability, predict performance, or replace professional advice.

Current status: Nexum Prism is not generally available. No live Prism consumer is connected to the engine, broker, custodian, or real-market data source.
Features

Separate evidence for
separate decisions.

The planned Prism experience will preserve the boundaries already enforced by Nexum Engine instead of presenting every decision as though one explanation covers it all.

Forecast Contributions

TreeSHAP feature contributions explain one synthetic forecast-model output with a tested additivity tolerance.

Constraint-Aware Allocation

Optimization evidence discloses constraints, residuals, binding bounds, return and risk contributions, sensitivities, and counterfactuals.

Trade Reconciliation

Trade evidence shows buys, sells, positions, cash, target weights, turnover, and total-value reconciliation.

Recommendation Replay

A hash-linked bundle preserves the exact forecast, allocation, and trade envelopes and supports deterministic replay.

Four-Dimension CRI

Correctness, Faithfulness, Stability, and Constraint Compliance are reconstructed from named engineering checks.

Explicit Audit Status

Signed local audit events and Prufs delivery states distinguish a created event from a verified durable cloud receipt.

Explainability boundary: TreeSHAP explains the forecast model only. Allocation and trade evidence use their own typed artifacts; Prism does not claim “SHAP on every recommendation.”
How It Works

From portfolio context
to a reviewable bundle.

This is the planned consumer flow, grounded in the exact sequence the engine already verifies with synthetic inputs.

1. Establish context and constraints
A future Prism consumer will collect authenticated account context, investor preferences, exclusions, liquidity needs, and risk constraints. This collection layer is not yet implemented.
Consumer
Planned
2. Generate sealed evidence
The engine creates the forecast artifact, constrained allocation, trade explanation, and hash-linked recommendation bundle in a fixed sequence.
Engine
Implemented on synthetic data
3. Reconstruct reliability checks
The CRI artifact evaluates named gates across Correctness, Faithfulness, Stability, and Constraint Compliance. Its score measures those gates, not investment truth.
Evidence
Implemented boundary
4. Present for human review
The planned interface will keep model contributions, allocation evidence, trade effects, limitations, and audit status visible before any user decision.
Experience
Planned
Preview Pricing

A proposed structure,
not an active subscription.

These target tiers are product-planning signals for preview participants. Checkout, billing, and paid account access are not available.

Individual
$29
target monthly price
Planned core portfolio evidence experience.
Advanced
$59
target monthly price
Planned expanded constraints and research context.
Professional
$99
target monthly price
Planned multi-account and professional workflows.

Pricing, features, and availability may change before launch. Joining the preview list creates no purchase commitment.

About

Built to make the boundary
between evidence and judgment visible.

Nexum Prism is part of CognitionHive, the product portfolio founded by Wade Lovell around measurable reliability, inspectable evidence, and human authority.

Wade Lovell

Wade's AI work spans econometrics, enterprise architecture, machine learning, and agentic systems. Across banking, insurance, and enterprise software, he has worked where technical decisions must remain explainable to operators, executives, and control functions.

His doctoral research develops the Claim Reliability Index around four dimensions: Correctness, Faithfulness, Stability, and Constraint Compliance. Nexum translates that concern for reliability into typed, replayable portfolio evidence.

PhD CandidateMS, Artificial IntelligenceColumbia MBACPAPublished Author
Preview List

Help shape an investor experience
that shows its evidence.

Join the list for product research and preview updates. This is not an account application, investment service enrollment, or offer to transact.