Frequently Asked Questions — Private & Offline AI

Plain-English answers to the questions our clients ask most about running AI on their own terms: securely, privately, and on their own premises.

Your data and your privacy

Where is our data stored? Entirely on your own machine, on your own premises. Nothing is sent to a third-party cloud, an external service, or any server we control. Your data never leaves your building unless you choose to send it.

Can you see our data? No. The system is built so that your operational information — documents, records, conversations — stays sealed within your environment. Where we retain a connection for maintenance, that access is technically limited to system functions only and contractually prohibited from touching your data. You hold the keys, and you can verify it.

Is our data encrypted? Yes. Information is encrypted at rest on the machine, and any maintenance connection we use is encrypted in transit.

Who owns the hardware and the data? You do — both. The machine sits on your premises and everything on it is yours. We provide, configure, and maintain the system; we never take ownership of what runs through it.

What happens to our data if we end our agreement? It stays with you. Because everything lives on your hardware, there is no data of yours held on our systems to return or delete. If you part ways with us, your information simply remains exactly where it always was — with you.

Offline and on-premises operation

Can the system run completely offline, with no internet connection? Yes. For organisations with the strictest privacy or regulatory requirements, we can deploy a fully air-gapped system — one that never connects to the internet at all. The AI models run locally on your machine, so the system works entirely within your own four walls.

Does running offline limit what the AI can do? For the great majority of tasks, no. Modern local models are remarkably capable and run comfortably on a single dedicated machine. The only trade-off in an offline setup is that updates are delivered manually rather than automatically — a small price for complete isolation.

Why choose on-premises over a cloud AI service? Control, privacy, and peace of mind. With a cloud service, your data is processed on someone else's infrastructure under their terms. On-premises, your information never leaves your control, you are not exposed to a third party's data policies or outages, and you can demonstrate exactly where your data lives — which matters enormously in regulated sectors.

Where does the AI processing actually happen? On the machine in your office. Every response and every action is processed locally. There is no hidden round-trip to an external server.

Access, maintenance and support

If the system is on our premises, how do you maintain and improve it? Through a secure, encrypted connection that you control — or, for fully offline clients, through scheduled on-site or offline updates. This keeps the system running smoothly and lets us apply improvements without requiring any technical knowledge from your team.

How do we know you're not accessing our data during maintenance? Three safeguards work together. First, our access is technically restricted to system functions: we can restart and update services without the ability to read your data. Second, every maintenance session is logged. Third, it is written into your contract — we are explicitly prohibited from accessing your data, and you may audit that connection at any time.

Can we revoke your access at any time? Yes, instantly, and on your terms. Withdrawing our access does not affect the system's ability to keep running. You are never dependent on us to remain operational.

How are updates delivered if we're offline? For air-gapped systems, updates are prepared by us, verified, and applied either on-site or via secure offline media, on a schedule that suits you. You receive the latest improvements without ever opening the machine to the internet.

What happens if the hardware fails? We design every deployment with recovery in mind. Your configuration and models can be restored to replacement hardware quickly, and where guaranteed uptime matters we can build in redundancy. We agree the right level of resilience with you before deployment.

Compliance and governance

Does this support our GDPR and UK data protection obligations? It is built for exactly this. Keeping personal data on your own premises, under your sole control, with no transfer to third parties, directly supports the data-minimisation, security, and residency principles at the heart of UK GDPR. We work alongside your team to ensure the deployment fits your specific obligations.

Will you sign a Data Processing Agreement? Yes. Where we act as a processor, we provide a clear Data Processing Agreement setting out precisely what we can and cannot do, your rights, and our obligations. For fully offline deployments our role is often more limited still — which we are always happy to document.

Can we audit the system? Yes. You can audit our access, review activity logs, and inspect how the system behaves. Transparency is a feature, not an afterthought: we would always rather you could verify our claims than simply take our word for them.

Is there a record of what the system and its agents do? Yes. The system keeps activity logs of the actions its agents take, giving you a clear, reviewable trail of what happened and when — supporting both your governance requirements and your own internal oversight.

Is this suitable for regulated sectors such as legal, healthcare, or finance? It is designed precisely for them. Organisations with professional, regulatory, or contractual duties of confidentiality are exactly who benefit most from a private, on-premises, optionally offline system — because it lets them adopt AI without surrendering control of sensitive information.

Getting started and commercials

Do we have to configure anything technical ourselves? No. We handle the entire setup. The machine arrives — or is installed — ready to work. You switch it on; we have done the rest. There is nothing for your team to install, configure, or troubleshoot.

Will it integrate with our existing tools? In most cases, yes. We assess your current systems during planning and build the integration to suit. Where strict isolation is required, we can equally run the system entirely standalone.

Are we locked in? No. You own the hardware and your data, so you are never trapped. Our aim is to earn your continued business through the value we deliver — not by holding your data or systems hostage.

What ongoing costs are involved? Typically a one-off setup alongside the hardware, then a maintenance arrangement suited to your needs. Because the system runs on your own machine, you avoid the per-use and per-seat cloud fees that climb as you grow. We set out the full picture transparently before you commit.Write your text here...