From change signal to verified action in three phases.
The Verification Graph Engine takes any proposed change and computes the full picture: what's impacted, how severe it is, what verification work is needed, and what it will cost.
Back to Impact Intelligence overviewChoose traversal scope
You control which parts of your operational graph get analyzed. Run everything, or focus on specific domains:
- Full traversal: all providers, every edge type
- Domain-scoped: target specific areas like supply chain, documents, or compliance
- Custom: combine any providers and edge types for precise queries
Provider domains
Domain providers span supply chain, compliance and quality, operations, documentation, and engineering, each traversing its area of the operational graph independently.
Select all providers for a complete analysis, or scope to specific domains for faster, role-relevant results.
Execute the pipeline
Seven steps transform a seed change into scored impacts, proof paths, verification work, and cost estimates.
Seed
A change request identifies what's changing and targets specific entities:
- Change types: engineering, process, software, asset, service, compliance
- Target entities: SKUs, BOM revisions, documents, drawings, routings, and more
Traverse
Domain providers walk the dependency graph across:
- Documents, workflows, and partner acknowledgements
- Composition structures and task execution chains
- Compliance requirements and verification checkpoints
- Capital assets and geometric / 3D data
- Procurement pipelines and inventory relationships
Enrich
Once traversal is complete, every impacted node is enriched with the context your team needs to act on it:
- Human-readable context: SKU names, document titles, work order numbers, status labels
- Permission-filtered: each user sees only what they're authorized to see
- Performance-optimized: fast even when traversals return hundreds of nodes
Score
Each impacted node receives a severity score (0.0 to 1.0) based on operational context. Scoring parameters are tenant-configurable.
Detect
Multiple changes happening at once can silently conflict. This step detects where their impact areas overlap:
- Serious conflicts are flagged for human review, automatically
- Built-in resolution workflow for triaging and resolving overlapping changes
- Prevents conflicting changes from silently undermining each other
Explain
Proof paths trace exactly why each node is impacted:
- Minimal hop-by-hop trace from seed to impacted node
- Deterministic and reproducible: same inputs always produce the same path
- Audit-ready: answer "why did we re-validate this?" for regulators and customers
Act
Verification packs generate the work needed to execute the change safely:
- Draft validation plans, document packages, acknowledgement tasks, operation checklists
- Configurable templates: a configurable rule engine drives artifact generation
- Preview mode shows all proposed actions before persisting anything
Why your provider selection matters
The providers you include in Phase I shape every output that follows.
The output of each run depends on what the traversal finds. Change which providers run, and all of these change with it:
- Severity scoring and hotspot distribution
- Cross-change collision detection
- Proof paths and audit evidence
- Verification packs and cost estimates
Full traversal
42 nodes
all providers
Complete change impact. Every provider fires. Use for final approval decisions and compliance evidence.
Supply chain providers
12 nodes
BOM + procurement providers only
Procurement team's view. Faster, focused on supplier certs and material impact. Skip work orders, CAD, docs.
Compliance providers
8 nodes
compliance + quality gate providers only
Auditor's view. Only regulatory and quality impacts. Generates targeted verification packs for re-validation.
Same seed change. Same seven-step pipeline. But a different provider selection yields different outputs:
- Different node count and severity distribution
- Different verification packs and cost estimates
- Fewer providers means faster, cheaper queries. Teams run impact checks more often
- Results relevant to your role, not buried in noise from unrelated domains