Inside your own
environment.

Lyrion is an engineering firm. We design and build AI systems inside the businesses that use them, on your own platforms and under your own controls, and we operate them after they go live. We work across insurance, construction, energy and financial services, and the same engineers scope, build and run each system.

Inside your own environment: your cloud tenancy or your own data centre, in the region you choose. The system runs within the platforms your teams already use, and nothing is migrated or replaced to make it work.

You do. Residency requirements are recorded at the scoping session with your security and legal teams, written into the blueprint, and signed off before anything is built. A change of region goes through change control like any other change.

No. A system reads your records where they already live and writes its results back to your own systems of record. Nothing is copied out to us and nothing is uploaded to a third party to make it work. Logs, backups and the audit trail stay in the same region as the data they describe. Where a model provider is used, it is named in the blueprint with the region it runs in and the terms it runs under, and a request that cannot stay in the required region is held for a person to decide. Your security team sees every one of those decisions in the audit trail, and any of them can be reversed by a person with the authority to do so.

No. Records, documents and conversations are used to do the work and nothing else. No model is trained or tuned on your data unless your team asks for it in writing and approves the scope, and model providers are held to the same terms in their contracts with you and with us.

Only the people and systems your team approves. Each system holds the access its process needs and nothing more, scoped per system. Credentials stay in your vault and keys stay under your management. Our engineers work inside your environment under the access you grant, every session is logged, and your team can revoke that access at any time. When a person needs to see a record to decide an exception, the system shows them that record and nothing else, and the view itself is logged against the record. Nobody at Lyrion browses your data.

Encrypted in transit and at rest, scoped per system, and held to your existing security policies. Your engineers review every connection before it carries real work, and every action the system takes is logged against the record it touched. Approvals sit wherever your team wants them, and nothing leaves the building without a person. That is what gets a system through a security review, and it is the standard we hold every system to after go-live.

It stays where it always was. Everything built inside your environment stays there, with its documentation, its configuration and its audit trail, so your team can keep running it, hand it to another provider, or switch it off. Nothing has to be returned, because nothing was ever copied out. The blueprint, the connections and the rules the system follows are written down from the first day, so nothing about the system depends on a person who has left. We prefer to stay, and we operate what we build, but the choice is yours.

Where residency requires it, yes. The system, its logs and its audit trail can all run inside your own data centre, and so can the models where the work allows it. Where a hosted model is needed, it is named in the blueprint and reached through a connection your engineers review and control, in the region you choose. Nothing about the system changes for the people who use it: it runs inside the same platforms either way.

Because that is where the work is. The process, the exceptions and the systems of record already live there, and so do the controls your auditors trust. A system built anywhere else has to be trusted twice: once for what it does, and once for where it does it. Building inside your environment means the system answers to your controls from the first day, and our engineers stay accountable for it after go-live. Nothing is migrated, nothing is replaced, and nothing is copied out to us to make it work. It is also why the same engineers scope it, build it and run it.

From the scoping session onwards. Your security team sees the written blueprint before anything is built: every connection, every credential, every data flow, and the region each one runs in. They review each connection before it carries real work, hold the approvals they want to hold, and receive the audit trail in the format their tools already read. When a rule, a connector or a model changes, the change goes through change control with the approver named on it. Nothing reaches production without their sign-off, and nothing changes in production without it either.

They can ask our engineers anything, at any point, and the answer is written down. Every question and answer is kept with the blueprint, so the next reviewer starts from what the last one learned. Our engineers stay on the same system for as long as it runs, so the people who answer your security team are the people who built it, and they know where every record goes.

Yes. Lyrion keeps every action logged against the record it touched, with the rule it followed and the person who approved it, inside your own environment. Your auditors can read the trail directly, without asking us for an export. Lyrion's commitments on data and residency are set out on this page, under Our commitment on data.

Yes. We operate what we build, inside your environment and under your controls, and extend it to the next process the same way. The same engineers scope it, build it and run it, so there is no handover to a team that was not in the room. When something breaks in production it is ours until it is fixed, and every fix is logged against the record it touched like any other action. The rules the system follows change only through change control, with the approver named on each change, and your team can read the full history of a system in one place.