Glossary · On-premise

What does on-premise mean?

Definition: On-premise software is installed and run on a company's own servers, rather than in the cloud on an external provider's infrastructure.

In an on-premise installation, the company hosts the application on its own infrastructure, whether in its own data centre or on servers it controls, and its IT team manages access, backups and updates. In the cloud model, the application runs on a provider's infrastructure and is accessed over the internet, often as a service. Strictly speaking, the correct form is on-premises, though both are widely used.

The choice depends on the company's security policy, its obligations regarding data and the technical resources it has available. Hosting in-house gives more control over where the data sits and who can access it. A provider's cloud frees the company from running servers and makes it easier to scale. Both options are valid if they are properly documented.

There are options in between: a private cloud that the company itself contracts and controls, or hybrid models that combine local systems with external services. In AI projects the question is particularly relevant, because the documents sent to the model may contain personal or confidential data. Neither option is inherently more secure; what counts is the quality of the controls.

Example: a company whose policy requires customer documentation to stay within its own infrastructure installs the AI assistant and the model on its own servers instead of using a cloud service.

Let’s talk about your processes.

A first conversation to get to know your company and see where we can help.