The Sovereign AI Platform
The AI advantage.
None of the exposure.
Frontier AI for security-conscious organisations.
From confidential chat to a powerful agentic analytics platform.
Get in touch
Why Cortex exists
The AI transition is here.
Harness it with confidence.
When you send a query to a commercial AI provider, you're telling them who you are, what you're working on, and how you think. Your prompts are logged, your identity is visible, and your data is retained under their legal obligations, not yours. For organisations handling sensitive information, that exposure is unacceptable.
The productivity gap widens every day. The organisations using AI effectively are compressing weeks of work into hours. For those who can't use commercial platforms, that gap is growing.
The biggest productivity gains aren't in chat. They come from a powerful analytics platform working across all your data. Commercial platforms can't reach it without exposing it. Sovereign alternatives need forward-deployed engineers. We built a third way.
We believe no organisation should have to choose between the benefits of AI and the confidentiality of their work.
By design, not by policy.
Sovereignty. Assurance. Safety.
Every design decision we make is oriented around three properties: sovereignty, assurance, and safety. Deliver all three and you have a platform that security-conscious organisations can use without compromise.
Sovereignty
Sensitive organisations should be able to keep their data, identities, and queries inside their own legal and technical boundary. We build products that let them do that. At the far end, that means running entirely inside client infrastructure. But even on shared infrastructure, Cortex removes any persistent link between organisations and their queries. No logs, no identity exposure, no metadata trails.
Assurance
Assertions about security are only meaningful when they can be tested. We give clients, auditors, and regulators the evidence they need to verify our claims: independent penetration testing, certifications such as Cyber Essentials Plus, and transparent architecture documentation. Assurance is the bridge between what we say and what we can prove.
Safety
Deploying AI inside sensitive environments means accepting responsibility for what it produces. Cortex gives organisations the tools they need to define and enforce their own AI usage policies: semantic query gating, classification-appropriate guardrails, and audit trails for every interaction. Safety is governance made operational.
One platform. Three products.
Built for the environments where security is not optional.
Arc
Familiar AI chat and document intelligence. Confidential by design.
Vantage
Agentic analytics. From dataset to decision.
Palisade
Sovereign AI aligned to classification requirements. For OFFICIAL-SENSITIVE and above.
ARCFamiliar AI. Confidential by design.
Arc delivers frontier AI through a familiar interface. Chat, document intelligence and multi-language support, with confidentiality and attribution protection built in at every layer.
- Shared filesets, team projects, and audit logs
- Confidential AI chat across all frontier models
- Document intelligence: query, synthesise and surface insights from your files
- Multi-language support across 90+ languages
- Cast-iron enterprise controls and admin oversight
- Semantic query gating, configurable to your rules and evaluated locally
Arc is complete out of the box. Scale and compliance requirements drive the choice, not access to capability.
Explore Arc & VantageVANTAGEProductivity through agentic analytics. Sovereign by design.
Vantage is the agentic analytics platform. Not a dashboard. Not a reporting tool. An AI layer that reasons across your structured data, builds pipelines, and produces outputs you can act on.
Most organisations sit on vast amounts of structured data they barely touch. The analysis is too slow, the tools too rigid, the expertise too scarce. Vantage changes that. Ingest datasets. Define goals. Let agents do the work.
Data stays inside your boundary. The same deployment model as Arc, from anonymised cloud to fully sovereign. No forward-deployed engineers. No external dependencies.
Public sector & defence
Cortex delivers frontier AI capability to UK government and defence, across classification levels. Alongside frontier AI chat and document intelligence, Cortex has developed specialist capabilities for government knowledge management, and for defence: capabilities to accelerate planning cycles, enrich HQ staff processes and tools for information operations.
Specialist capabilities
- Government knowledge management
- Accelerated planning cycles
- HQ staff process enrichment
- Information operations tooling
- Sub-Palantir economics for civil departments via G-Cloud
We welcome enquiries from government and defence organisations and their supply chains.
Deployment
Built for your security posture.
One platform. Three ways to deploy it, each built for a different security posture.
Anonymised
Queries routed through Cortex infrastructure with full attribution protection. Your organisation is never identifiable as the source.
Enclave
AI processing runs inside a dedicated private cloud environment, including assured government hosting. Queries never reach AI company systems.
Sovereign
Deployed on client-owned infrastructure or a private network. No external dependencies. Air-gapped capable.
Our guarantees
Security as architecture.
Not policy.
No attribution
Your organisation can never be identified as the source of a query. By design, not policy.
No training. Ever.
Your prompts are logged and aggregated. Your data is never used to train or improve any model. Guaranteed contractually.
Your deployment, your rules
From shared sovereign cloud to fully air-gapped. The architecture matches your requirements, not the other way around.