Protect your investment. De-risk your AI future.
Independent, court-defensible AI risk assessments using the Shadow Simplex scoring methodology. Designed for deal teams, boards, and GRC officers who need more than a checkbox.
Our flagship engagement produces a Shadow Simplex Score (SSS) report — a systematic, reproducible risk assessment derived from a patent-pending geometric framework. Two offering tiers are available:
Seven-factor defeasibility-weighted SSS across the dimensions enumerated below — XSC, SCS, ECI, MDI, SAI, CCI, VII — with shadow-simplex centroid analysis and tier classification. Coverage under US Provisional 64/066,231.
Adds the five-register stratification, the capability normalizer C(κ), the 9 × 6 risk-primitive matrix, and the transcendental meta-condition veto layer. Non-compensatory multiplicative aggregation throughout. Coverage under SSPLX-002-PROV.
Unlike qualitative audits or vendor-supplied model cards, the SSS report is:
Delivered in 2–4 weeks. Includes executive presentation + remediation playbook.
We review your AI system inventory, deal timeline, regulatory context, and key risk areas. You leave with a clear sense of engagement scope, timeline, and cost range — usually within 24 hours of the call.
We issue a scoped proposal with fixed pricing, clear deliverables, and timeline commitments. NDA and engagement letter signed before any data sharing begins.
We conduct the full SSS assessment: documentation review, architecture interviews, technical testing, regulatory mapping, and shadow AI discovery. You receive a weekly status update with any emerging findings flagged immediately if material.
Draft report delivered for factual accuracy review. We don't negotiate findings — but we correct factual errors and incorporate context you provide. Final SSS score does not change after this stage without written agreement.
Final report package delivered: written report, SSS scorecard, executive presentation deck, and 30/60/90-day remediation roadmap. We present findings to your board, deal team, or GRC committee — included in all engagements.
Every dimension is assessed independently using diagnostic protocols derived from the Shadow Simplex Framework, then weighted to produce the composite score.
Measures whether AI systems across your stack produce consistent, coherent outputs — or whether multi-system pipelines introduce compounding errors and semantic drift. Tests for cascade failure conditions.
Evaluates the governance controls in place: model lifecycle management, access controls, output monitoring, human-in-the-loop configurations, and incident response capability.
Maps the system's current posture against applicable regulatory obligations — EU AI Act Article 9/17/69, NIST AI RMF, SEC cybersecurity disclosure rules, HIPAA, DORA, and SR 11-7 as applicable.
Assesses the organization's ability to detect and respond to model drift — both conceptual drift (changing input distributions) and behavioral drift (shifting output patterns over time). Identifies monitoring gaps.
Discovers and catalogs AI systems operating outside formal IT governance — including personal LLM accounts, unregistered API integrations, and departmental deployments without enterprise review. Often the highest-risk dimension for large enterprises.
Models failure propagation risk: if one AI system in your architecture makes a significant error, how far does it propagate? Identifies architectural chokepoints and missing isolation boundaries.
Reviews training data provenance, copyright exposure in model outputs, IP ownership structures for custom models, and the AIBOM (AI Bill of Materials) for third-party components. Critical for M&A transactions where AI is a material asset.
No commitment, no NDA required for the initial call. We'll tell you what a scoped engagement looks like for your specific situation.
Book scoping call