We spent three decades watching the same systems fail in identical ways.
Legacy PLM, ERP, MES, CMRO, and EAM were designed to manage data and enforce process. They were never designed to model reality, reason across lifecycle boundaries, or tell you whether a decision is trustworthy. Clarity was built from scratch to do what they cannot. Every capability on this page is a direct answer to a failure pattern we documented across defence, aerospace, nuclear, automotive, and space programmes over thirty years.
The pattern we documented
Same failures. Every programme. Every decade.
The marketing changes. The architecture doesn’t. These are the six failure patterns we observed repeatedly across PLM, ERP, MES, CMRO, and EAM — and the structural answer Clarity builds for each one.
Legacy claim
“Digital Thread”
Seamless lifecycle traceability from design to service. End-to-end visibility. One coherent model.
What actually happens
IDs copied between systems via nightly sync jobs. PLM has the eBOM. ERP has the mBOM. MES has production. MRO has maintenance. The “thread” is a reconciliation spreadsheet refreshed weekly. Ask which system is authoritative and watch the meeting go sideways.
What Clarity builds instead
13 lifecycle layers from stakeholder intent (L0) to disposal (L12). Schema-validated provenance at every node. Every entity has a lineage chain. @source is part of the data model, not an audit log bolt-on.
Legacy claim
“Change Management”
Structured approval workflows. Governance enforced. Audit trail maintained.
What actually happens
Bureaucratic approval chains that take weeks for simple changes. Urgent fixes bypass the system entirely. A massive audit trail records who clicked Approve — but has no idea why the change was needed, what it impacted, or what decision authorised the programme of work it belongs to.
What Clarity builds instead
L5 Decisions are the structural parent of change, baseline, release, and configuration management. A change is valid because an L5 decision approved it. Change management is a consequence of decision-making — not a parallel discipline nobody trusts.
Legacy claim
“Configuration Management”
The system reflects what is physically installed. Configuration controlled. One version of truth.
What actually happens
Ask “what is actually installed?” and the honest answer is “depends.” Configurations are dynamic and networked. Legacy systems use relational joins and status tables. Reality diverges. Excel becomes the source of truth within eighteen months of go-live.
What Clarity builds instead
16 BOM views on one CI graph. Every configuration item exists once. Versions, variants, alternates, and substitutes are explicit graph relationships with qualification evidence — not parallel trees or shadow items. DeZolve truth vector validates every claim before it becomes your source of truth.
Legacy claim
“AI-Powered Insights”
Intelligent automation. Predictive analytics. AI-generated recommendations across the enterprise.
What actually happens
Document summarisation and recommendation heuristics layered onto systems with broken semantics, poor relationships, and bad data quality. AI cannot compensate for incoherent data models. Incoherent inputs at enterprise scale — just faster. Hallucinations presented as analysis.
What Clarity builds instead
AI that generates structured model entities, not prose. 15 years of DeZolve decision intelligence and a proprietary world model constrain the LLM to the schema. The lessons-learnt knowledge graph flags what has gone wrong before. Intelligence is in the platform — not the model.
Legacy claim
“Integrated Platform”
One coherent enterprise suite. Seamless interoperability. Shared data model across all modules.
What actually happens
Tight coupling sold as simplification. PLM “material” ≠ ERP “part” ≠ MES “item.” Point-to-point interfaces that fail under load. Integration teams appear. Reconciliation meetings recur weekly. Adding a new module requires an integration project and a consultant. Upgrades become impossible.
What Clarity builds instead
Diode and airlock connectors that exchange structured data without forcing standardisation. Three independent enforcement layers — not configurable away by application logic. Your existing tools continue doing what they do; Clarity provides the schema-validated CI graph between them.
Legacy claim
“Best Practices Built In”
Adopt proven process templates. Industry best practices delivered out of the box. No reinvention required.
What actually happens
Adapt your business to the software. “Best practice” means easiest for the vendor to standardise and support — not best for your programme. Organisations distort their operations, lose flexibility, and pay consultants to customise anyway. The customisation accumulates into a fork that can’t be upgraded.
What Clarity builds instead
Open JSON. Customer-owned KMS keys. Any standard is a configuration overlay, not a module. You define your process; Clarity enforces the structural constraints (ISO 10007, CMII) and adapts to your workflow — not the other way around. Your data is readable without a Clarity licence in 10 or 50 years.
Patent pending • First published 2012 • Silicon Valley
The DeZolve Decision Intelligence Framework
Decision intelligence is not a feature Clarity added. It is the core of what Clarity is. DeZolve was developed and first published in Silicon Valley in 2012, validated across defence, nuclear, automotive, and major technology programmes over a decade, and is now the patent pending engine at the centre of the Clarity platform. No commercial tool at any price point implements anything equivalent.
Problem Classification
DeZolve classifies every decision as Tame (well-defined, solvable with standard methods) or increasingly Wicked (emergent constraints, contradictory stakeholders, no objectively correct solution). Each class triggers a different analysis pipeline with appropriate confidence bounds. Legacy tools treat every decision as a form to fill in. DeZolve treats each one as the problem it actually is.
The Truth Vector
For every formal decision, DeZolve traverses the Lx chain — explicit traceability links, AI-inferred connections, and gap assessment — and computes a trustworthiness score from live evidence before you commit. Not assembled from a stale report the night before a gate review. Not a traffic-light dashboard of user-asserted status. A structural computation from the evidence chain that exists at the moment of decision.
Evidence Coverage Panel
Every L5 Decision Record displays a coverage map of the Lx chain: which layers have formal evidence (green), partial evidence (amber), or gaps (red). The Decider sees exactly what they are and are not deciding on the basis of. Every legacy toolchain leaves this question unanswered until the audit.
Invariant Enforcement
Design invariants — inviolable programme constraints extracted from the ConOps by AI — act as standing guards. Any proposed change that would violate an invariant is blocked at the API level before it reaches the approval workflow. No PLM, requirements tool, or change management system does this. The constraint is embedded in the model, not in a process document nobody reads.
BOM truth validation uses the same DeZolve pattern. Every imported BOM claim — quantity, effectivity, usage — is authority-weighted before promotion. Evidence sources are ranked: as-built records (1.0), ECR/ECN history (0.95), deployment data (0.85), MES records (0.80), ERP orders (0.75). Contested items block promotion until a resolution rationale is recorded with full audit trail. This is the answer to the auditor’s question no PLM vendor can provide: not just “what does the BOM say?” but “what physical evidence supports it, and who resolved the conflicts?”
AI that knows what it doesn’t know
Intelligence in the platform, not the model
Every competitor’s “AI-powered” offering is a language model connected to a document store or a database. The intelligence is in the model — which means it hallucinates, it drifts, and it can’t be audited. Clarity’s intelligence is in the platform.
Three layers constrain every AI generation to the engineering model: the Lx schema (typed structure), the DeZolve decision intelligence framework (decision context), and the proprietary world model (MBSE ontology and knowledge graph built over 15 years). The lessons-learnt knowledge graph adds a fourth layer: AI-extracted signals from what has gone wrong in previous programmes, surfaced as structured flags on newly generated entities.
BYOM: bring your own model. Any LLM, including on-premises via SageMaker. Commercial AI APIs are not available in most classified AWS regions. Every competitor’s AI stops at the classification boundary. Clarity’s doesn’t.
Proprietary World Model
A complete MBSE entity taxonomy and knowledge graph built over 15 years — systems, components, functions, interfaces, requirements, capabilities, activities, constraints and their typed relationships. The AI extracts, classifies, and links against this model. Structural hallucinations are impossible when generation must produce entities the model recognises.
Lessons-Learnt Knowledge Graph
Structured capture of what has gone wrong — failure modes, programme risks, decision patterns that led to cost blowout or schedule slip — extracted from decades of programme experience. When AI generates a new entity, the knowledge graph signals whether analogous decisions have been contested or reversed before. Institutional memory that doesn’t retire.
BYOM — Any LLM, Any Environment
AWS Bedrock (Nova, Claude, Titan), on-premises SageMaker, or any API-compatible model. The intelligence layer is model-agnostic: the schema, DeZolve framework, and world model do the constraining. Swap the model without reconfiguring the platform. Deploy in air-gapped environments where no commercial AI API is available.
Six structural answers
What Clarity actually delivers
Each capability below is a direct structural answer to a failure pattern. Not a feature. Not a module. A different way of modelling the problem.
Replaces: nightly ETL jobs pretending to be lifecycle intelligence
13-layer Lx chain — the real digital thread
DOORS stops at requirements. Teamcenter covers BOMs. No legacy tool spans the full lifecycle. Clarity’s Lx chain is 13 layers from stakeholder intent (L0) to disposal (L12) — each layer with typed inputs, outputs, and bidirectional traceability links to every other layer. L0–L5 live at beta. L6–L12 post-beta. Non-linear entry at any layer: most programmes don’t start at L0, and Clarity imposes full structural discipline from wherever the programme actually is.
+ L6 As‑Designed • L7 As‑Built • L8 As‑Validated • L9 As‑Deployed • L10 As‑Operated • L11 As‑Updated • L12 As‑Disposed (post-beta)
Replaces: bureaucratic approval chains that substitute workflow for engineering understanding
L5 is the structural parent of change, baseline, release, and configuration management
In every legacy toolchain, change management (PLM module), baselines (database snapshots), releases (separate pipeline), and configuration management (separate discipline) are disconnected — and none reference the decision that authorised them. In Clarity, all four are structural children of the L5 Decision record. A baseline exists because an L5 decision froze it. A change is valid because an L5 decision approved it. Configuration management is the totality of L5 decisions + L4 baselines + the CI graph — not a parallel discipline maintained in parallel with the model.
Cross-tool change cascade: GitHub PRs, Teamcenter ECRs/ECNs, Jira Issues, SAP approvals, CVE findings — all mapped to their CI-graph dependencies. A firmware PR that triggers a hardware ECR that requires an ERP substitute approval is visible as a cascade before the change is approved — not three months after it has been built into out-of-configuration hardware.
Replaces: reconciliation spreadsheets, phantom assemblies, and shadow BOMs
16 BOM views. One CI graph. Zero shadow items.
Teamcenter’s tree architecture was designed for CAD file management. Variants require parallel trees. ECAD and software don’t fit. Bulk items (fasteners, wire, consumables) become phantom assemblies requiring manual reconciliation. Clarity sidesteps the tree entirely: every configuration item exists once in the graph; all 16 BOM views — from eBOM through decommissioning — are live graph queries. Three acquisition modes (as-designed, as-acquired, as-inherited) handle the real-world complexity of multi-supplier programmes without separate tools.
The reconciliation spreadsheet replaced by a quality-gated pipeline. BOM imports from any source (PLM eBOM, CAD tree, ERP mBOM, flat parts list, CIDL) pass through a six-stage Library pipeline: ingest → parse → standard-part auto-tagging → DeZolve truth validation → PBS editor review → promote to L6. Every claim is authority-weighted before promotion. Contested items block promotion until resolution is recorded. Multi-supplier BOMs are harmonised via deterministic CI identity per supplier namespace. The six-figure PLM integration project and the weekly reconciliation meeting are replaced by a standard Library workflow.
Replaces: slow, wrong, stale reports that require three weeks and a consultant to extend
Documents that write themselves — computed, not assembled
Legacy reports are expensive screenshots of bad data. Generated nightly. Semantically wrong because the underlying data model is incoherent. Not queryable dynamically. The answer to “show me the evidence trail” is “give us three weeks.”
Clarity’s scope-matrix report engine covers 13 lifecycle phases × 39 assessment layers. The number of valid report scopes is combinatorially unbounded — not a fixed catalogue requiring professional services to extend. Every report is a computed projection of the live Lx chain: always current, always faithful to the underlying data. When an auditor asks “show me the traceability from requirement 3.2.1 to your verification test, against which baseline, who approved the change, and what decision it led to” — the answer is a link, generated in seconds.
33 report types covering ISO 15288, ISO 10007, and MIL-STD-973. Domain packs add sector-specific report templates for aerospace (AS9100), medical (ISO 13485), defence (MIL-STD-810), automotive (ISO 26262), rail (EN 50126), and nuclear (NQA-1).
Replaces: point-to-point spaghetti, semantic mismatches, reconciliation teams
Diode & Airlock connectors — structured exchange without forced standardisation
Clarity does not force standardisation on any vendor stack. Your PLM continues managing CAD. Your ERP continues managing procurement. Clarity provides the schema-validated CI graph that connects them — via diode (one-way ingest, write-back structurally impossible) or airlock (two-way, policy-governed, fail-closed on infrastructure failure) connectors.
Three independent enforcement layers — IAM inline deny, S3 bucket policy, no reverse event rule — are not configurable away by application logic and are independently auditable without reading code. For AUKUS and Five Eyes programmes: cross-classification digital thread sharing across AU/UK/US programme boundaries. CrossDomainTraceLink bilateral audit trail. Deployable from a single CDK stack definition and a JSON configuration file.
For SMBs and Tier 2–4 suppliers without enterprise tooling: Clarity is a direct replacement for the full PLM, PDM, MRO, and MBSE stack — from AU$750 / £400 / US$500 per user per year, browser-based, no infrastructure, no systems integrator, hours to first deliverable.
Replaces: vendor lock-in, proprietary formats, and data that becomes unreadable after contract expiry
Your data. Your keys. Readable without a licence in 10 or 50 years.
Every programme entity is stored as documented open JSON. No proprietary format. No export required to read it. No Clarity licence required to read it in 10 or 50 years. Per-tenant KMS keys — owned by the customer, inaccessible to Clarity employees by architecture. Data residency controlled at deploy time by AWS region selection.
For regulated programmes with long life cycles: S3 Object Lock (WORM), versioning, and daily manifest-driven backups ensure nothing is silently overwritten and any programme state is reconstructable at any point in time. Deploy to your own AWS account and Clarity has zero dependency on the operating instance. For nuclear, defence, and long-life aerospace programmes: the 30-year question “why was that decision made?” is structurally answerable without institutional memory, without the original tooling, without the original team.
Who it’s for
Every entry point leads to the same platform.
Builder — SE lead, SMB engineer
Bidding for contracts that require MBSE artefacts you can’t produce without outsourcing
Full Lx chain in hours, not months. AI generates the first-pass model from a ConOps. Scope-matrix reports for bid and audit submission. No SysML specialists, no systems integrator, no infrastructure. From AU$750 / £400 / US$500 per user per year.
Governor — PM, auditor, quality, compliance
Facing a regulatory audit with a folder of Word documents as your evidence trail
Structured evidence chain, ISO 10007 change management, requirement-to-decision trace, fully reportable. Domain standards loaded as overlays: AS9100, ISO 13485/FDA, NQA-1, ISO 26262, EN 50126. When the auditor asks, the answer is a link — not a three-week assembly exercise.
Integrator — prime, multi-tier, multi-nation
Connecting DOORS, Teamcenter, SAP, and twelve supplier BOMs without a reconciliation meeting
Clarity as the schema-validated CI graph between your existing tools. No rip-and-replace. For AUKUS and Five Eyes programmes: cross-classification digital thread exchange via diode and airlock patterns. For multi-tier supply chains: DeZolve truth validation on every supplier BOM claim before promotion.
Guardian — security, compliance, sovereignty
Deploying engineering intelligence in an environment where commercial AI APIs are not cleared
BYOM: any LLM including on-premises via SageMaker. Deploy to your own AWS account, air-gapped regions, or ATO-partner-managed infrastructure. Customer-owned KMS keys. ITAR/EAR/NOFORN enforced by three independent structural layers — not configurable away by application logic.
Free tier available: L0–L5, 3 users, 1 product, 3 GB — self-service at app.compliancewithclarity.com. No credit card required. Full pricing →
One thread. 13 verticals. 16 BOMs. 25 USPs.
The only complete digital thread for regulated programmes, powered by the patent pending DeZolve Decision Intelligence Framework. Sovereign deployment under your own AWS account and encryption keys — at 10× less than the enterprise alternatives.