Operational Domains
Where Periphery operates
One platform. Four verticals. The same emergent architecture adapts to each domain because it doesn't impose structure — it discovers it.
Supply Chain Intelligence
The Problem
Global supply chains span hundreds of jurisdictions, thousands of entities, and constant regulatory change. Sanctions lists update daily. Shell companies obscure beneficial ownership. A single missed connection can mean regulatory exposure or operational disruption.
How Periphery Solves It
Periphery ingests sanctions lists, corporate registries, shipping data, and news feeds simultaneously. The Crystallizer resolves entities across languages and jurisdictions — the same company known by three names in three countries collapses into one node. Emerging clusters surface chokepoint risks before they become crises.
Example Workflow
An analyst investigating a supplier receives an alert: the Crystallizer has identified a new cluster linking their Tier 2 supplier to an entity on the OFAC SDN list through a shared beneficial owner in a third jurisdiction. The connection was invisible in any single data source. The Continuous Critic assigns the cluster a "Defined" legibility tier — high enough coherence for immediate review, with full source provenance attached.
Financial Compliance
The Problem
AML teams process thousands of alerts daily. False positive rates exceed 95%. Beneficial ownership networks span multiple registries and jurisdictions. Regulatory requirements change faster than manual processes can adapt.
How Periphery Solves It
Periphery's five-space embedding architecture captures patterns that rule-based systems miss. The relational embedding space maps beneficial ownership networks. The temporal space detects behavioral patterns — transaction cadence, registration timing, filing gaps. The Continuous Critic eliminates false positives by demanding coherence across all five spaces.
Example Workflow
The system identifies a cluster of corporate entities with temporally synchronized registration dates, shared registered agents across three jurisdictions, and transaction patterns that mirror known layering techniques. The Critic scores the cluster "Emerging" — coherent enough to warrant analyst attention but with explicit uncertainty markers on two of the five entity resolutions. The analyst reviews, confirms the entities, and the cluster upgrades to "Solid."
Geopolitical Analysis
The Problem
Geopolitical events generate signal across dozens of source types — diplomatic cables, social media, satellite imagery metadata, economic indicators, local news. Analysts drown in volume. Traditional tools require manual correlation. Critical signals hide in the noise of routine reporting.
How Periphery Solves It
Periphery's geospatial and temporal embedding spaces capture the spatial and chronological signatures of geopolitical events. The semantic space identifies thematic connections across languages. Force movements, diplomatic signals, and economic indicators cluster automatically — revealing patterns that would take human analysts weeks to identify.
Example Workflow
Periphery detects an emerging cluster: increased military logistics activity in a border region, correlated with unusual diplomatic communications patterns and local media reporting on infrastructure changes. Each data point alone is routine. Together, they form a cluster the Critic scores as "Emerging" with a specific confidence breakdown across each embedding space. The analyst can see exactly which dimensions drive the assessment and which remain uncertain.
Cyber Threat Intelligence
The Problem
APT campaigns reuse infrastructure, modify tooling incrementally, and operate across months. CVE feeds, malware samples, infrastructure scans, and dark web monitoring generate massive data volumes. Connecting a new indicator to an existing campaign requires correlating across all of these sources simultaneously.
How Periphery Solves It
Periphery treats cyber threat data as another domain for its five-space architecture. Infrastructure overlaps surface in the relational space. Malware evolution tracks in the semantic space. Campaign timelines emerge from the temporal space. The Crystallizer identifies campaign clusters that span traditional CTI boundaries.
Example Workflow
A new malware sample triggers enrichment through multiple feeds. Periphery's Crystallizer identifies infrastructure overlap with a previously dormant cluster from six months ago. The relational embedding connects shared C2 patterns. The Critic scores the connection "Haze" — possible but requiring additional indicators. When a second indicator confirms the infrastructure link two days later, the cluster coherence jumps to "Defined" and the analyst receives a prioritized alert with full campaign timeline.
Your domain. Your data. Emerging structure.
Periphery adapts to the intelligence problem. Not the other way around.
Request Demo