Cybanetix unveils Managed AI Service to secure users, models, and agents

Cybanetix has announced the launch of its Managed AI Service to address all three aspects of AI use within the enterprise. Covering employee AI usage, AI governance, and embedded AI, the Managed AI Service combines technology from NOMA, SentinelOne, Microsoft, and Exabeam with Cybanetix consultancy, managed services, and 24/7 Security Operations Centre (SOC) monitoring.

The result is a managed service for AI security within the enterprise, capable of responding to alerts in under 15 minutes.

AI is being deployed faster than it can be secured and the risk falls across three domains. The first is the user behaviour, individual employees uploading data into public large language models (LLMs) or using unsanctioned models. The second is the governance, including model provenance, AI bill of materials (AI-BOM) and ownership of AI assets across the business. The third is embedded AI, where agents and tools are wired into business processes, often with excessive privileges and no clear inventory.

Each domain has different controls, vendors and specialists compelling the organisation to buy point solutions which inevitably create gaps in the security posture.

The Managed AI Service provides a 360-degree view of the AI security ecosystem and includes observability and exposure mapping, behavioural monitoring of AI activity across the estate, runtime protection at both the infrastructure and application layers, and synthetic and adversarial testing of the models themselves.

SentinelOne Prompt Security and Microsoft Purview for AI handle user-level controls. NOMA delivers AI discovery, access control, red teaming, and detection and response, with findings mapped to ISO 42001, the EU AI Act, and the NIST AI RMF. Exabeam covers agent behaviour analytics.

The wrap around the stack is provided via Cybanetix services and its 24/7 SOC operations. The consultancy arm covers AI security policy and strategy, AI security deployment and integration, and posture analysis as well as the hardening of systems.

The SOC service provides AI security platform management, AI observability and real-time threat detection for AI specific risks such as prompt abuse, model manipulation or anomalous AI behaviour. In concert with the technology stack, the Managed AI Service therefore not only detects AI security events but also correlates, responds and improves the security posture through the delivery of:

• AI Risk Assessment – Discovers and builds inventory of every AI component in use; maps agent-to-agent relationships and provides a visual agentic risk map detailing the blast radius of each integration; reports on AI risks; identifies control gaps and recommends actions.
• Technology deployment – Utilises the SentinelOne Prompt, NOMA suite of services, Microsoft Purview for AI and Exabeam UEBA/ABA solution to enable observability and provide runtime protections aligned with AI usage policies.
• AI posture management – Triages and prioritises risks; enforces policies across no/low-code and custom-built AI environments, protecting internally developed AI and safeguarding training data from misuse or compromise; platform management supports changes or configurations made during ‘business as usual’; provides integration with SOC playbooks and Continuous Exposure Management (CEM) dashboards.
• Managed Detection and Response (MDR) – Integrates with wider SOC/MDR functions and correlation with identity and Endpoint Detection and Response (EDR) solutions to enable risk triage and escalation; automates containment using options from blocking to intercepting traffic or notifications sent to the end customer coupled with human-led response that includes full incident forensics and executive reporting.
• Continuous AI Risk Reporting – Delivers visibility of risks and exposures to track AI posture improvements and assess and demonstrate AI usage against security, compliance and governance requirements; provides CEM AI reporting; maps findings to compliance standards ISO27001, SOC2, NIS2 and Cyber Essentials Plus and aligns with frameworks including the EU AI Act, ISO42001, NIST AI RMF.
• Testing and validation – Includes numerous tests from real-time detection of prompt injection, anomalous behaviours and data exfiltration to AI red teaming to continuously surface exploitable weaknesses.

The Managed AI Service confers numerous benefits over and above those of any standalone solution. Events are integrated with workflows and automation, making detection and response faster, and because they are correlated with EDR and identity contextual data can be integrated with wider SOC/MDR investigations and response. This means the organisation gains full visibility of the AI posture alongside other attacks. The service continuously discovers and controls AI tools, agents and models and customers also gain the benefit of access to AI experts, preventing the need to hire inhouse AI security specialists.

“AI tools such as co-pilots, coding agents and LLM-powered workflows are now embedded across every business function and represent the fastest growing, least-protected entry point in the enterprise. CISO’s need AI posture management, model provenance, red teaming for rogue behaviour, real-time runtime detection, and the evidence trail to satisfy regulators and auditors. No single product does that, whereas the Managed AI Service covers all those bases to provide a truly credible AI security programme,” explains Merlin Gillespie, CTO at Cybanetix.