Glossary
Compliance terms, defined.
Short, plain-language definitions for every term used across Clawscan — product concepts, regulatory frameworks, and Microsoft 365 deployment terms. Cross-linked so you can follow a thread in any direction.
Clawscan product terms
Concepts specific to how Clawscan works.
Credit
- The unit of measurement for Clawscan usage. One credit equals one email scanned against one active legal domain. An email scanned against three active domains consumes three credits.
- RelatedLegal domain
Guard
- Guard is Clawscan's reactive compliance module. Every email routed to your Clawscan shared mailbox is automatically scanned against all active legal domains, with a timestamped verdict stored for audit. 100% coverage, not sampling — unlike manual compliance review. Guard is your second line of safety: even when something slips through, your Legal & Compliance teams are the first to know.
- RelatedShieldVerdictLegal domain
Intent-based detection
- Clawscan's detection approach. Rather than matching banned keywords, Clawscan's AI models evaluate the intent of an email against regulatory risk patterns — e.g. recognising price coordination whether or not the word 'price' appears. Reduces false positives compared to legacy email DLP keyword filters.
- RelatedLegal domainDLPVerdict
Justification
- A plain-language explanation attached to every Clawscan verdict — what the AI detected, which regulatory framework applies, and why escalation is warranted. Unlike keyword alerts, a justification gives a compliance officer or general counsel actionable evidence instead of a raw flag.
- RelatedVerdictIntent-based detection
Legal domain
- A regulatory risk area covered by its own purpose-built Clawscan AI model. Current domains include Competition Law, Conflict of Interest, Confidential Information & Trade Secrets, and Workplace Harassment, with more added continuously. Each domain can be activated independently; credits are only consumed for active domains.
- RelatedIntent-based detectionCredit
Privacy filtering
- A client-controlled Clawscan feature that skips scanning for emails whose subject line matches a configured keyword (e.g. [Private], [Confidential]). Matching emails are dismissed before entering the scan pipeline — never read, never stored, never processed. The keyword list is managed entirely by the client.
- RelatedTenant-local architectureGuard
Shield
- Shield is Clawscan's proactive compliance module. It runs inside Outlook while an employee is drafting and surfaces a verdict — classification and written justification — before the email is sent. Designed to prevent legal & compliance risk from ever leaving your Microsoft 365 tenant.
- RelatedGuardVerdictCredit
Tenant-local architecture
- Clawscan's deployment model in which raw email content is scanned inside your own Azure tenant and never crosses the tenant boundary. Only scan outputs — classification and justification — are transmitted to GOlegal for storage. Minimises data egress and aligns with GDPR data-minimisation principles.
- RelatedMicrosoft 365 tenantDLPGDPR Article 5
Verdict
- The structured output of a Clawscan scan. Every verdict contains two components: a classification (critical / warning / safe) and a written justification referencing the specific flagged passage. Verdicts are what make Clawscan output audit-defensible rather than a black-box signal.
- RelatedJustificationShieldGuard
Regulatory & compliance terms
Legal frameworks and enforcement concepts referenced across Clawscan.
Article 101 TFEU
- The core EU competition law provision prohibiting agreements between undertakings that restrict competition — including cartels, price coordination, bid-rigging, and market allocation. Violations can trigger fines of up to 10% of global annual turnover. Email is routinely the primary evidence source in EU cartel investigations.
- RelatedPer se offenceLegal domain
DLP
- Data Loss Prevention — a category of security software that blocks the transmission of predefined sensitive data (credit card numbers, SSNs, classified keywords). DLP is content-signature matching; Clawscan is intent-based legal risk detection. DLP answers 'is this data banned from leaving?'; Clawscan answers 'does this communication describe unlawful conduct?'
- RelatedIntent-based detectionTenant-local architecture
FCPA
- Foreign Corrupt Practices Act — US federal law prohibiting bribery of foreign public officials and imposing accounting controls on issuers. Along with the UK Bribery Act, one of the two principal anti-bribery frameworks with extraterritorial reach.
- RelatedLegal domain
GDPR Article 5
- The GDPR provision establishing data processing principles, including purpose limitation, data minimisation, and storage limitation. Article 5 underpins Clawscan's tenant-local architecture: raw email never leaves the controller's tenant, so processing is minimised to the scan outputs actually needed for the compliance purpose.
- RelatedTenant-local architectureLegal domain
MAR
- Market Abuse Regulation (EU Regulation 596/2014). The core EU framework governing insider trading and market manipulation in regulated markets. Covered firms must detect, investigate, and report suspected market abuse — including communications evidence.
- RelatedLegal domain
Per se offence
- A category of regulatory violation that is unlawful by its nature, regardless of actual market effect. Most hardcore cartel conduct — price fixing, bid-rigging, market allocation — is per se unlawful under EU and US competition law. An email agreeing to coordinate prices is an offence whether or not prices were actually aligned.
- RelatedArticle 101 TFEU
Microsoft 365 & technical terms
Platform, architecture, and deployment terms relevant to Clawscan deployments.
Dataverse
- Microsoft's managed relational data store used by Clawscan to persist scan outputs. Encryption at rest uses TDE / AES-256. Dataverse provides the audit-ready evidence store that sits behind the Guard dashboard.
- RelatedTenant-local architectureGuard
Entra ID security group
- A Microsoft Entra ID (formerly Azure AD) group used to define which employees are in scope for Clawscan. A single security group controls both the Shield add-in deployment and the Guard shared-mailbox redirect. Dynamic groups are supported for attribute-based scoping (role, department, location).
- RelatedMicrosoft 365 tenantShieldGuard
Microsoft 365 tenant
- A customer-dedicated instance of Microsoft 365 / Azure, identified by a tenant GUID. All Clawscan scanning happens inside the client's own tenant — Shield runs as an Outlook add-in in-tenant; Guard uses an in-tenant shared mailbox. No tenant boundary is crossed by raw email content.
- RelatedTenant-local architectureEntra ID security group
Outlook add-in
- The task-pane component that runs Clawscan Shield inside the user's Outlook client (Windows, Mac, Web, Mobile). The add-in is deployed centrally through Microsoft 365 Admin Center and requires no per-user install. It activates for employees inside the configured Entra security group.
- RelatedShieldEntra ID security group
See these terms in practice.
Book a 30-minute demo and see how Clawscan works for your organisation.