Designed for accountability
The ICO's accountability framework rewards systems with named processors, documented purposes, and records that can be produced on request.
Self-hosted AI / the UK
Private AI deployed for British businesses under the UK GDPR and the Data Protection Act 2018. Hosting in a London region, a documented data path, and a model provider that is named before anything is loaded.
Example only / architecture varies by client
Why it matters here
ICO guidance and a monitoring-focused employment practice code make disclosure the practical bar.
The ICO's accountability framework rewards systems with named processors, documented purposes, and records that can be produced on request.
ICO's code on monitoring at work is part of the review environment for staff-facing systems, and the deployment's data path is written for exactly that review.
eu-west-2 (London) keeps the document store close, with the option to stay fully local when contracts demand it.
The three levels
There is no single correct answer. Quality, privacy, residency, and budget decide the path.
App and document store in your account in eu-west-2 (London), with the model on your key. No GPU required.
An approved cloud region or managed service when residency and procurement matter.
Model, embeddings, and retrieval on your hardware when external inference is not acceptable.
What it really takes
Hardware is sized after the workload is real. These are starting points, not quotes.
No GPU required for the app and document store. Model usage bills stay on the client account.
A starting point for a small quantized text model. Model size, concurrency, and latency decide the real requirement.
A GPU is one part of the environment. A partner data centre can be more realistic than an office server room.
We size hardware after the workload is proven. No GPU inventory before a signed scope and deposit.
Straight answers
The default is eu-west-2 (London). UK GDPR and DPA 2018 considerations are part of the documented design.
No. The ICO does not pre-approve systems. The deployment is designed to be reviewable — named processors, documented purposes, access rules, and a runbook.
Where systems process staff data, ICO's monitoring guidance applies. The documented data path is built for that conversation with HR and employment counsel.
Yes. The ballpark starts at £2,900 for a first workflow, with USD and naira floors unchanged.
Why trust Pristine3D?
Pristine3D Ltd builds and operates live digital products, and we run private AI workflows internally as part of our own operations. We scope around your real workflow: the documents you own, the questions your team asks, and the access boundary you approve. Based in Lagos, Nigeria, we work remotely with clients worldwide.
Based in Lagos, Nigeria, Pristine3D currently builds and operates smartcards.ng, venu.ng, photoshoot.ng, and ugc.ng in production.
One input, one output, one test set, and one person who owns the result. We scope a real workflow instead of a transformation programme.
Cloud, model, storage, and messaging accounts stay in your name. The chosen data path, access rules, test record, documentation, and training are part of the agreed scope.
Pricing / fixed scope
The final quote follows the workflow. Infrastructure and model bills stay on your accounts.
One workflow for a small team, with training and handover.
See the one-workflow packageMultiple sources, roles, integrations, and admin handover.
See the deploymentStandard care for one delivered workflow. Optional. Larger deployments and active monitoring are separately scoped.
See supportArchitecture review from $500. Standard annual support is $3,000 / ₦1.5m per year for one delivered workflow. New workflows, integrations, active monitoring, and infrastructure are separately scoped; infrastructure, model, storage, and messaging bills stay on client accounts. Full pricing and what changes the quote
Own your knowledge base
Documents, the retrieval index, access rules, and the workflows built around them are the asset, and they compound. We deploy so the knowledge base stays yours: on your accounts, in the environment you choose, under access rules your team defines. The model behind the answers is a connector, so the knowledge base moves with you, not with a vendor.
The document store, metadata, and retrieval setup live on accounts you own. No vendor holds the corpus.
Change the model provider, move regions, or go local without rebuilding the knowledge base or the workflow.
Every improvement to the corpus improves the answers, and the improvement stays with you, not with a vendor.
Keep exploring
Start with the constraint
Residency, procurement, staff-data rules, or a no-external-inference requirement. Describe it and we will map the architecture.