Skip to main content

AI & Threat Intelligence

This area does two different jobs that share one idea — bring context to a finding so a person spends less time deciding and more time fixing. Threat Intelligence brings the outside world in (what is being exploited, by whom, where it maps on ATT&CK) and pushes it into prioritisation. The AI pages turn the platform's own data into decisions: triage verdicts, prepared fixes, an executive briefing, answers from your policies — and, because you run AI too, a registry and posture for your AI systems.

Where: left navigation → Threat & IntelligenceThreat Intelligence and AI Governance; Core SecuritySecurity Command Center; ManagementKnowledge Base.

What is in this section

PageWhat it isRead it for
Threat IntelligenceNine curated feeds (CISA KEV, URLhaus, OTX, Feodo Tracker, SSL Blacklist, PhishTank, Blocklist.de, Spamhaus DROP, OpenPhish) normalised into indicators, hourly; IOC correlation; CVE prioritisation; MITRE ATT&CK coverage; landscape reports; STIX / CSV import-exportWhich of your exposures are tied to active exploitation
Security Command CenterTriage agent, remediation queue, Auto-Fix Engine with approvals, Security Advisor, activity log — rule-based out of the box, LLM-assisted when a provider is configuredWhat to fix first and why, with an audit trail
AI GovernanceModel registry, risk assessments and bias tests, incidents and human-oversight logs, training records, ISO 42001 control posture and auditor exportsGoverning the AI systems your organisation builds and buys
AI Discovery, AIBOM & TestingDiscovery of AI services in your cloud inventory and LLM configuration; the AI bill of materials folded from your SBOMs; OWASP LLM01 prompt-injection testsKnowing what AI you actually run, in code and in cloud
Knowledge Base & AI AssistantPolicy library with sections and sensitivity, question answering with citations, questionnaire auto-fill with an answer bank, usage analyticsAnswering "what is our policy on…" and security questionnaires from your own documents
AI Data & PrivacyWhich features call a model, which never do, what leaves the platform, caching and tenant isolationThe commitments behind every AI feature
API · Troubleshooting & FAQEndpoints and permissions; common problems

How the pieces connect

  1. Feeds → indicators → priority. Every hour the enabled feeds are fetched and normalised into indicators (IPs, domains, URLs, hashes, CVEs, certificates). A CVE that appears in CISA KEV, or is tied to active indicators, actors or campaigns, scores higher in Vuln Prioritization and in the Triage engine's exploitability signal.
  2. Findings → triage → fixes. The Command Center's triage agent classifies open findings into Critical / Investigate / Monitor / False positive / Auto-resolved using severity, KEV, EPSS, exposure, environment and age. Findings that match a remediation playbook become Auto-Fix actions that wait for approval; code findings with a fix disposition can be handed to the agentic fix pipeline.
  3. Your AI → registry → posture. Discovery finds AI services in the asset inventory and LLM configuration; the AIBOM finds AI libraries in your SBOMs; both feed the model registry, whose assessments, tests, incidents and training records roll up into the ISO 42001 control posture.
  4. Your documents → answers. Knowledge Base documents are chunked and embedded; the assistant and the questionnaire filler answer from them with citations and a confidence score, and approved answers build an answer bank.
With and without an AI provider

Threat intelligence, rule-based triage, the Auto-Fix Engine, the Security Advisor's data-driven answers, discovery, the AIBOM and the governance registry all work with no model provider configured. An LLM (Anthropic, OpenAI or Google — configured per team under Knowledge Base → AI Assistant → AI Configuration) adds reasoning to triage, FP/TP assessment, remediation drafting, the briefing, and document Q&A; Knowledge Base retrieval specifically needs an OpenAI key for embeddings. Each page says which mode it is describing. See AI Assistant for provider setup.

Permissions at a glance

ActionPermission
Read the threat dashboard, indicators, feeds, heatmap and reportsany authenticated member
Correlate, prioritise, generate a landscape report, configure or refresh feeds, add indicators, actors, campaigns, rules, aging, import / exportManage Threat Intelligence (manage_threat_intelligence)
Run agents, triage, FP/TP assessment, ask the advisorRun AI Agent (run_ai_agent)
Approve / deny / execute / roll back Auto-Fix actionsExecute Remediations
AI governance registry, assessments, incidents, oversight, discovery, prompt testsManage Assessments
Knowledge Base uploads, questions, questionnaire fillManage Assessments; reading the library and analytics needs only a login

Everything is scoped to the active team: feeds, indicators, agents, registry entries and documents belong to the team that created them.