Evidence & Uploads
Manage Microsoft security evidence, assessment coverage and data quality. As at 31 Oct 2026 · compared with 30 Sept 2026
How Evidence Coverage is calculated
A control is fully evidenced only when every dataset it requires is present, has a valid status, is within its freshness period and covers its full scope.
Partially evidenced controls (19, 18%) are reported separately and never added to coverage.
Four layers — kept separate
- 1 · Evidence availability1 · Availability23/29 datasets
- 2 · Evidence freshness2 · Freshness22/23 current
- 3 · Control implementation3 · ImplementationNot assessed here
- 4 · Control effectiveness4 · EffectivenessNot assessed here
Coverage reflects layers 1–2 only. Implementation and effectiveness are assessed separately in the Material Risks process.
Single source of truth
This calculation is the only source of Evidence Coverage. The Executive Overview KPI, its trend and insights, and the Cyber Risk Index (15% evidence-gap weight) all use the same value.
Evidence catalogue versions
This assessment uses v1.0 (29 datasets · 106 controls). v2.0 was approved 12 Nov 2026 and applies from the Nov 2026 assessment — historical definitions are preserved. v3.0 (endpoint evidence, 38 datasets · 108 controls) was approved 13 Nov 2026 and applies from the Dec 2026 assessment.
For Oct 2026, v2.0 would report 61% against the applied 62%. The new control (IAM-15) raises the denominator to 107, and controls that now require sign-in, Identity Protection, PIM or Defender for Identity evidence cannot be fully evidenced until those datasets are collected.
Most remapped controls were already not fully evidenced under v1.0, so the effect is at most one point of coverage and no change to the Cyber Risk Index (15% evidence-gap weight, rounded). The Executive Overview continues to report the applied v1.0 values for these assessments.
Evidence upload centre DEMO · SIMULATED
Files stay in this browser tab. Nothing is uploaded, stored or analysed.
What happens in this demonstration
- 1Select files locallyOnly name, size and type are read. Contents are never opened.
- 2Simulated transferProgress is animated in the browser; nothing is sent to a server.
- 3Name & format checkExtension, size limit and export pattern are checked.
- 4Proposed classificationCategory and confidence from the filename — for analyst review.
- 5Schema validationPlanned: parse real CSV, JSON and PDF exports.
- 6Secure storage & processingPlanned: backend ingestion engine.
Session queue is empty — add files or try a sample export.
Microsoft export recognition
Proposed classification from filename and extension only
Identified export patterns (11)
→ category- M365_SecureScore_RecommendedActions01
- Azure_SecureScore_Recommendations_HighCritical01
- ConditionalAccess02
- Endpoint_IntuneDeviceComplianceStatus09
- Endpoint_VulnerabilityList_AllCVEs03
- Endpoint_PatchRemediationRecommendations03
- DefenderForCloud_Recommendations_FullExport_AllSeverities04
- Azure_InternetFacing_PublicIPInventory06
- DefenderForCloud_RegulatoryCompliance_AzureCSPM04
- DefenderO365_PhishDetections05
- DefenderO365_ContentMalwareDetections05
Evidence categories
12 categories · 29 required datasets · 106 mapped controls · catalogue v1.0 · status as at 31 Oct 2026
Evidence gaps
10 required datasets missing or incomplete · prioritised by impact on assessability
Evidence history
19 of 84 files received to 31 Oct 2026 · select a file for metadata and validity
RiskLens · Demonstration build · Evidence, files and validation outcomes shown are fictional. No files are uploaded or stored.