About
Built capability, not a concept deck
AI is being deployed and written faster than assurance can scale. AICAP is built capability on production-grade foundations, from a team with a delivery record in the environments where this matters most.
Why now
AI is outpacing assurance. Systems proliferate faster than manual review can follow.
AI coding went mainstream. Output has jumped an order of magnitude; review capacity has not.
Provenance is broken. Accept an AI suggestion and the record of which tool and model wrote it is gone.
The bar is rising. Customers, auditors and emerging standards increasingly expect provable assurance.
Why Digital Native Group
Already built. Working pipelines, a validated correlation engine, LLM infrastructure and a running product.
Built for regulated industry. Financial services, healthcare, energy and critical infrastructure carry the same burden as government: show the system is safe, keep showing it, and hand the working to someone who will check. That is where most of the demand sits, and it is what the platform is shaped around.
Cleared and trusted. An SC-cleared team with secure delivery across Cabinet Office, FCDO, MoJ and MoD. The standards that clear you for that work, evidence, traceability and careful handling, are the same ones a financial services or healthcare auditor asks for.
Sovereign by design. Lattice, our air-gapped Kubernetes distribution, on open foundations. Full freedom to operate.
Open where it builds momentum. A published provenance schema as a candidate standard, with the value in the platform.
How you run it
One build, three deployment models. Which one you pick is a question about where your data is allowed to live, not about what you give up.
Managed
We host and run it. Quickest to get started, and the right choice when your data is allowed to leave your estate.
Your cloud
Deployed into your own tenancy, under your controls and your key management. You keep the data and the boundary.
On-premise & air-gapped
No external connectivity, on Lattice, our air-gapped Kubernetes distribution, with self-hosted open-weight models so inference never leaves the estate.
Who we'd like to work with
We are bringing in a small number of people to take the work further.
Engineers & builders
Backend, data pipelines, LLM integration or secure infrastructure, and interested in shaping a product early.
AI assurance & security minds
Depth in model evaluation, adversarial testing or code security, to help define how this works in practice.
Design & validation partners
Organisations willing to test the platform against real systems and give honest, early feedback.
Aligned backers
People who understand the direction and want to help it move faster than it could alone.