Private AI for
your business.Set up and
managed for you.
Your company’s information, tools, and approved workflows. A controlled AI environment. A provider responsible for setup, operation, and support.
A guide from Dominus Foundry, creators of Sabina.
Your business. Your boundaries.
Choose the information, people, and permitted actions. Agree what stays outside the service.
A model to discuss with a provider. Not a product demonstration.
Company context.
Clear responsibilities.
Start with the work your business needs to do. A managed service brings together selected company information, connected tools, and rules about who may do what. The provider’s responsibilities continue after the initial setup.
“Private” needs a concrete explanation: where information travels, who can access it, how long it stays, and which uses are allowed. Those details belong in the service agreement and the demonstration.
Here, managed AI means a service with agreed responsibility for deployment, operation, maintenance, and improvement. Managed private AI adds explicit controls over business information, processing, access, and use.
“What did we promise customers this week, and what still needs attention?”
This is a walkthrough of a possible workflow, using hypothetical records. It is not a customer result or a claim about a currently demonstrated Sabina capability.
- 1
Start with permitted records
Use only the project notes, commitments, and messages the requester is allowed to see.
- 2
Show the supporting source
List each commitment with its record, due date, owner, and any uncertainty that needs checking.
- 3
Propose a follow-up
Prepare a draft for the responsible person. Keep an incomplete or conflicting record visible.
- 4
Respect the approval
Send or update only within the agreed permission. Record what actually happened, including failure.
What the provider should manage
A useful proposal names the work, the owner, and the evidence that it is being done.
- Setup & connections
- Configure the environment, agree information boundaries, and maintain the connections used in the approved work.
- Access & changes
- Handle joiners, role changes, departures, and changes to which actions require approval.
- Operation & support
- Monitor the agreed service, investigate failures, and name a support contact and response expectations.
- Improvement & exit
- Test changes against representative work. Agree what can be exported and how access is removed when the service ends.
Privacy should be something you can inspect.
Ask for a data-flow explanation, a permission demonstration, and clear terms for retention and outside services.
Explore privacy and control →Is a managed service the right fit?
For occasional drafting, a simpler subscription may be enough. For coordinated work across company records and tools, compare the value of ongoing management with the work your team can own.
Compare the responsibilities →
Meet Sabina.
Dominus Foundry publishes this guide and develops Sabina. Explore the product’s current information and assessment to discuss the work your business needs.
Explore Sabina →