AI Control Centre

Pilot available

Control your AI before it controls you. One place every AI application registers with, takes its brand and security rules from, and reports into. The live dashboard is designed to show what is running, who is using it and what it costs.

Talk to us Ask for the preview
Every AI points to one placeIllustrative
AI Control Centre register · guide · report Support assistant Legal summariser Meeting notes bot Code review agent Third-party tool New application Brand guidelines Security rules Reporting protocol Guidance out, through MCP Reports in

How it works

The starting and reporting point for all your AI, whether new, already running or built by someone else.

Register

Every AI application registers first, with an owner and a purpose. Third-party tools that cannot report are registered too, with the monitoring limit shown.

Get guidance

Before any building starts, the application pulls your brand guidelines, security rules and reporting protocol through MCP.

Report in

It reports its health, requests, errors, speed, tokens and estimated cost back to the Control Centre.

Watch live

The dashboard shows what is running, what is in use and what it costs, on a service desk screen or in the workspace.

Live AI dashboard

See the whole portfolio at a glance, with the problems at the top.

  • Health for every application, with incidents first
  • Usage by application and model, with dated weekly figures
  • Cost as an estimate, kept apart from what providers bill
  • TV display for the service desk, plus a read-only page for your intranet
  • Manage workspace for owners, guidance, reports and readiness
AI Control Centre, TV display
The AI Control Centre TV display showing application health, requests, failures and tokens by model, with illustrative demo data.
Design preview with illustrative demo data.
AI Control Centre, MCP connections
The MCP connections page listing brand assets, security and protocols, approved documents and a project registry.
Design preview with an example catalogue. No services are connected.

One set of rules for every AI

Applications retrieve what applies to them through MCP, so every new build starts from the same standards.

  • Brand guidelines and approved assets
  • Security rules and current protocols
  • Reporting protocol so every app reports the same way
  • Approved resources and tools, with versions and owners

Honest by design

We only report what can be shown. The Control Centre keeps three things apart that most reporting blurs together.

Said versus checked

DeclaredSomeone says it is fine
VerifiedA check shows it is fine

Estimated versus billed

EstimateOur figure from usage
ConfirmedWhat the provider charged

Silent versus healthy

No dataNothing reported
HealthyA passing check

Start with the method, grow into the platform

The framework (a 40-point checklist, maturity scoring and a retrofit method) works without any software, so you can start now.

Self-service

Framework pack

Checklist, scoring guide, retrofit method and templates.

With Kubatana

Guided start

Policy, register and guidance set up with you, plus one pilot application.

By Kubatana

Retrofit

Discovery, register, triage and a baseline score for AI already in use.

Ongoing

Control Centre

Regular reviews and reporting, with the live dashboard as it is built.

The dashboard shown is a design preview with illustrative data. The framework supports good practice. It is not legal advice, certification or a compliance guarantee.

How confident would you be?

If someone asked you today what AI is running in your organisation, could you show them? Start with a short scoping conversation.

Talk to us