Analysis Jobs

Overview#

An analyst kicks off an entity extraction job against a 50,000-word document corpus. It will take several minutes. She does not want to sit and wait, she wants to continue working and know when it finishes, how much it cost, and exactly what was found. The Analysis Jobs module provides that visibility.

It tracks asynchronous AI-powered analysis operations with detailed progress reporting, token usage accounting, cost estimation, and structured result aggregation. Every long-running job is observable from creation through completion, giving analysts and administrators full transparency into the analytical work the platform is performing on their behalf.

Key Features#

• Asynchronous Job Management: Track long-running AI analysis tasks from creation through completion with full lifecycle visibility.
• Progress Tracking: Real-time progress updates (0-100%) for active jobs, enabling frontend dashboards to display live status to analysts.
• Token and Cost Monitoring: Track tokens consumed and estimated costs per job, providing visibility into AI resource usage and enabling budget management.
• Multi-Tenant Support: Jobs are scoped to individual tenants for secure data isolation across organisations.
• Investigation Association: Link analysis jobs to specific investigations for contextual tracking and per-case resource attribution.
• Structured Findings: Capture job results in a structured format including entity counts, relationship counts, anomaly counts, and prediction counts.
• Aggregate Statistics: Dashboard-ready statistics including job counts by status, average duration, total token usage, and cumulative cost estimates.
• Programmable API Access: Full API support for querying jobs with filtering and pagination, and retrieving aggregate statistics.

Job Types#

• Entity Extraction: Extract entities such as people, organisations, locations, and financial instruments from text data.
• Graph Analysis: Analyse relationship graphs to identify network structures, communities, and central actors.
• Threat Modelling: Perform security threat analysis to identify vulnerabilities and attack vectors.
• Pattern Recognition: Identify recurring behavioural patterns across datasets.
• Anomaly Detection: Detect statistical outliers that may indicate fraud, insider threats, or unusual activity.
• Correlation Analysis: Correlate data from multiple sources to identify connected events and relationships.
• Predictive Modelling: Generate forecasts and predictions based on historical data and patterns.

Job Lifecycle#

1. Queued: Job created and waiting to be picked up by a worker process.
2. Running: Job actively processing with progress updates.
3. Completed: Job finished successfully with results stored.
4. Failed: Job encountered an error, with diagnostic message captured.
5. Cancelled: Job manually cancelled before completion.

Use Cases#

Intelligence analysts run multiple concurrent analysis jobs against collected evidence and monitor a live dashboard showing progress, estimated completion times, and preliminary findings as they emerge.

Financial crime investigation units track AI token usage and costs per investigation, allowing supervisors to understand the analytical depth applied to each case and manage budgets across active caseloads.

Fraud detection teams associate entity extraction and anomaly detection jobs with specific investigation records, creating a complete audit trail of the analytical work performed and the findings generated.

Platform administrators monitor job queue depth, failure rates, and processing times to ensure the analysis pipeline remains healthy during high-volume periods such as post-breach investigations.

Integration#

The Analysis Jobs module integrates with other Argus modules:

• Analysis: Provides the underlying analytical engines that create and execute analysis jobs.
• Analytics: Analysis jobs may be created by the analytics engine for scheduled or triggered calculations.
• Administration: Job queue health and statistics feed into administrative monitoring dashboards.
• Investigation Management: Jobs are associated with investigations for contextual result tracking.

Open Standards#

• GraphQL (June 2018 specification): The entire Analysis Jobs API is exposed as a typed GraphQL schema using Strawberry, with strongly typed enums for job status and type, cursor-style paginated connections, and structured input filters for querying jobs.
• JSON (RFC 8259): Job findings, per-job configuration, and metadata are stored and returned as JSON objects, making results portable and consumable by any conforming client.
• ISO 8601: All job lifecycle timestamps, created, started, and completed, are serialised as ISO 8601 date-time strings, ensuring unambiguous interchange across time zones.
• JSON Web Token (RFC 7519): Every resolver enforces authentication via RS256-signed JWTs issued by the platform's auth service; unauthenticated requests are rejected with HTTP 401 before any data is accessed.
• OAuth 2.0 Bearer Token (RFC 6750): The JWT credentials are transmitted and validated as OAuth 2.0 Bearer tokens, aligning with the platform-wide authorisation transport convention.
• Role-Based Access Control (NIST SP 800-192): Tenant-scoped permission checks are enforced on every query resolver, ensuring that an analyst in one organisation cannot read or enumerate jobs belonging to another tenant.

Last Reviewed: 2026-02-05 Last Updated: 2026-04-14