Our Work
A portfolio of AI governance and cybersecurity systems designed, built, and deployed by the Netrity team across regulated financial services, healthcare, critical infrastructure, and enterprise environments.
Every engagement we take on is grounded in the same principle: that AI systems in high-stakes environments must be explainable, auditable, and governable from the ground up. The projects below represent that principle in practice, across sectors where the cost of getting it wrong is real.
Etheras: AI Cybersecurity & Governance Platform
Sector
​​​​Cybersecurity · Enterprise
Status
Deployed
Etheras is Netrity's flagship platform conceived, designed, and built end to end by our founding team. It delivers real-time threat detection with explainability at its core: every alert, every automated decision, and every output is traceable and interpretable by the security teams who rely on it.
​
Bias correction operates continuously across detection outputs. Counterintelligence modules identify and adapt to emerging attack patterns. Integrated compliance auditing supports regulatory obligations under NIS2, DORA, and the EU AI Act, without burdening the security operations centre.
​
Governance is not a control layer added to Etheras. It is a structural property of the system.
Private LLM System for Regulated Fintech
A large language model deployed entirely within a regulated payments environment, isolated from the cardholder data perimeter, with no public egress. Designed from the ground up to meet PCI DSS and EU AI Act requirements.
​
A policy-enforcement gateway sits ahead of the model, handling identity validation, prompt-injection controls, data-loss prevention, and tool authorisation. Retrieval is confined to approved, access-controlled knowledge bases. Every consequential action requires explicit human approval. Every interaction is written to an immutable audit trail.
​
Built for an environment where a governance failure is a regulatory event.
Sector
Financial Services · Regulated AI
Status
Deployed
Agentic AI Oversight System
As autonomous AI agents move from experiment to enterprise deployment, the question of control becomes critical. This system sits between AI agents and the enterprise infrastructure they act on, intercepting every consequential action before it executes.
​
A deterministic policy engine evaluates each attempted action against the agent's delegated authority, data classification, jurisdiction, financial impact, and current security posture. Actions are allowed, restricted, routed for human approval, denied, or suspended. Approved actions execute under a single-use authorisation bound to the approved parameters, and the full decision and execution chain is recorded for audit.
​
The system also discovers unauthorised shadow AI operating across the organisation, a risk most enterprises do not yet have visibility into.
Sector
Enterprise AI · Regulated Environments
Status
Deployed
AI Governance Assessment Engine
A formal assessment engine that evaluates an AI system's governance readiness across four structural dimensions: algorithmic fairness, model transparency, data governance, and human oversight.
​
Scoring is non-compensatory, every dimension must clear its sector-specific threshold independently. A single weak area cannot be averaged away. Each control is mapped directly to the regulatory obligation behind it, across the EU AI Act, GDPR, ISO/IEC 42001, NIST AI RMF, and Swiss revFADP.
​
The output is a deployment determination and an executive action plan, designed to be presented to a board or regulator, not filed internally and forgotten.
A limitted version is available in the AI Goverance Decission Support System section as a demo here.
Sector
AI Governance · Regulatory Compliance
Status
Deployed
Explainable AI and Model-Audit System, Clinical Deployment
Designed and built end to end, from hardware to model, and validated in a live regulated clinical environment: a home-based rehabilitation platform guiding patient recovery and detecting 17 comorbidities in real time.
​
The system exposes the model's reasoning at every step, surfaces bias across outputs, and produces a traceable evidence record that satisfies clinical safety, GDPR, and MHRA requirements. Every decision the system makes can be interrogated by a clinician, an auditor, or a regulator.
​
This work established the explainability-first principle that runs through every system Netrity builds.
Sector
Digital Health · Regulated AI
Status
Deployed
Continuous Assurance & Regulatory Intelligence System
Most organisations treat compliance as a periodic event. This platform turns it into a continuous operational process.
​
Governed evidence pipelines ingest telemetry from across an organisation's estate, identity systems, SIEM, cloud infrastructure, payment systems, and AI registries. An assurance engine tests in real time whether controls are operating correctly, distinguishing a genuine control failure from missing or unreliable evidence. A regulatory intelligence engine monitors evolving legislation and supervisory guidance, maps each obligation to the affected systems and control owners, and surfaces what has changed and what action is required, before a regulator asks.
​
Currently at prototype stage for a regulated fintech client.
Sector
Fintech · Regulatory Compliance
Status
In Development
Aegis: AI Engineering & Assurance Platform
An integrated system that develops advanced AI capabilities and continuously proves they remain reliable, secure, explainable, and authorised for the operational environment they act in.
​
Algorithms are designed, challenged in simulation, and adversarially tested before any operational use. Each approved capability is issued a validated operating envelope, defining the tasks, data conditions, confidence thresholds, and human oversight requirements under which it may act. In operation, a control plane checks consequential outputs against that envelope and reduces authority automatically when assurance conditions are no longer met.
​
The system detects drift, adversarial attack, behavioural change, and explanation instability, triggering requalification rather than waiting for periodic review.
Currently in development for a private defence client.
Sector
Defence · Critical Infrastructure
Status
In Development