Start an AI pilot

Private AI that runs on your servers or your cloud

Use AI while keeping your data under your organization's control.

Private AI means running AI where your organization controls the data, on your own servers or in your own cloud account, instead of sending documents to a public AI service. BUG MAI can configure Private AI with access control and logs, and chooses the setup with you after reviewing your data in discovery.

DefinitionPrivate AI

Private AI: AI that runs under the organization's own control, for example on its own servers or in its own cloud account.

It is used when documents are sensitive or internal rules do not allow sending data to an outside service.

§ 01 — Steps

How it works

  1. 01

    Classify the data

    You provideWhich documents and systems the AI may touch
    You getA written data map, including what must never leave
  2. 02

    Choose where it runs

    You provideYour servers, cloud account or any constraints
    You getA recommended setup with the trade-offs explained
  3. 03

    Pilot inside your boundary

    You provideA sample dataset and people to try it
    You getResults measured against the target, with data staying where agreed
  4. 04

    Operate and decide

    You provideYour go or no-go
    You getAn operating plan if you go

§ 02 — Connects to

What it can connect to

What we can connect to depends on the exports or interfaces your systems offer. We check this during discovery.

  • On-premise servers
  • Your cloud account
  • Login and identity systems
  • Document stores
  • Existing apps

§ 03 — Safeguards

Data and control

Each item below is something that can be configured, as agreed in the scope.

Where it runs
Can be configured to run on-premise or in your own cloud account.
Access
Can be configured with role-based access to who can use the AI and which documents they can see.
Logs
Can be configured to record who asked what, and when.

Privacy notice

§ 04 — Compare

On-premise, your own cloud and public AI services

On-premise AI, AI in your own cloud and public AI services compared
TopicOn-premiseIn your own cloudPublic AI service
Where data is processedOn servers you ownIn a cloud account you controlIn the provider's service, possibly abroad
Who runs the hardwareYou supply and maintain serversThe cloud provider runs the hardwareThe AI provider
Best whenData must stay inside your building or networkYou want control of data without running serversDocuments are not sensitive and outside processing is acceptable

§ 06 — FAQ

Common questions

Do we need our own servers?

Not necessarily. Private AI can run in your own cloud account. We recommend a setup after looking at your data and constraints in discovery.

Is it as capable as public AI services?

The capability of models that run privately can differ from public services. We test on your own task in the pilot and tell you the limits plainly.

Does any part use an overseas provider?

If a design uses an overseas AI provider, we say so in writing before work starts, and you decide.

See all questions

§ 07 — Get started

Start with one task your team wants fixed