Centru

Centru runs your AI operating model.

Boards, regulators, customers, and acquirers are asking how you know your AI is under control. Centru.ai shows where your organization stands, helps you decide what’s worth building, and evidences that it works safely.

  • Build safely
  • Align strategy
  • Reduce risk
  • Evidence the work

Why Centru

Building AI solutions has never been faster. Knowing what’s worth building, and proving it’s safe, is the hard part.

Many AI pilots stall before production because it’s hard to see where you stand, defend where you invest, or evidence that what’s built is safe to run. Centru helps you build the evidence to avoid chaos, from the first question to the release decision.

Think it through

Ask about priorities, investments, or a decision, like which AI workload to fund next, and get answers grounded in the model, with the reasoning shown.

Decide where to invest

Assess your organization across twenty-five weighted dimensions, get tailored findings and next steps, and reassess to see how far you’ve come.

Run your strategy

Set your North Star, the goals that matter most to your organization, then build the roadmap and see what’s on course, what’s at risk, and what’s coming due.

See the risk you carry

Record every AI workload’s uses, risks, and safeguards, and test each risk with Microsoft ASSERT, so each release rests on evidence before a person signs it off.

Build what’s next

Centru Skills builds alongside the coding tools your teams already use, on nearly any technology with an SDK, at low cost and with the evidence recorded as it goes.

Who it’s for

Built for everyone driving AI.

One model across the whole organization, not a tool for a single team. Here’s the question Centru answers for the people who run AI.

CIOs, Chief AI Officers and AI leaders

“Where do we stand, and where should we invest?”

Know where you stand and where to invest, with a score you can defend, and give the board measurable evidence of progress.

CFOs

“Which AI is worth funding?”

See which AI is worth funding before you fund it, with adoption and outcomes tracked against your North Star, the goals that matter most to your organization.

Legal, compliance and AI governance

“Can we show our AI is governed?”

Technical evidence you can hand to counsel, auditors, and regulators, built as the work is done. Centru informs the decision; your people make it.

Engineering and AI operations

“How do we move faster, safely?”

Build fast without shipping what you can’t defend, with every release tested and its evidence attached.

Investors, acquirers and M&A

“What are we really buying?”

Diligence before the data room opens, on a target or supplier, measured against the same model.

AI Diligence

For any organization that wants its AI governed and defensible, and proven where the bar is highest: banking and financial services, law, healthcare, energy and utilities, government, and the UN system and intergovernmental organizations.

Proof points

Assurance, backed by evidence.

Centru at work with a global bank, and live in the room at the AI Trust Forum. Cloud Lighthouse also runs its own AI practice on Centru.

Global bank

2,500+

Tests generated by Microsoft ASSERT, built into Centru.ai, against one agent, measuring what it actually did.

AI Trust Forum New York

74 min

Idea to Azure. Centru mapped 14 risks against the EU AI Act and NIST for legal review, then Centru Skills built a cost-aware asylum triage prototype and deployed it to Azure.

AI Trust Forum London

90 min

Idea to Microsoft Purview. Subject access requests can cost up to £16,000 each to fulfil. Centru Skills wrote the spec and product plan, and a rough prototype followed.

Centru 3 and Centru.ai

The model, and the platform that runs it.

Centru 3 is the model: a published, weighted standard for running AI well across an organization. Centru.ai is the platform where your teams use it. It’s a model, not a large language model, built to work with the AI models you already trust.

Centru.ai The platform · five modules

  • DashboardSee the whole picture
  • ExpertiseAsk the model
  • MaturityKnow where you stand
  • StrategyKnow where to act
  • WorkloadsScale AI with confidence

Centru 3 The model

Centru Skills The agents

Agentic teams that build alongside the coding tools your teams already use, in your environment, under the direction of your people.

Centru IQ The accumulated knowledge

Patterns, practices, and research, like The Frontier Campus, that let the model reason about your situation rather than recite a checklist.

Centru Core The weighted foundation

Five pillars and twenty-five weighted dimensions, with the maturity model that measures your organization against them.

Centru Core

What a well-run AI program looks like.

Five pillars and twenty-five weighted dimensions, on a five-level maturity model. Every dimension is scored, so gaps can’t hide behind strengths.

1

Strategy and Vision

The aspirations, roadmap, and rigor behind the program.

  • Executive Vision
  • Actionable Roadmap
  • Ecosystem Map
  • Programmatic Rigor
  • Center for Enablement
2

Ecosystem Architecture

The technical estate on which AI depends.

  • Core Platform Services
  • Data Distribution
  • Integration
  • Business Applications
  • AI Development Tools
3

AI Workloads

The AI that advances the mission, from embedded to differential.

  • Workload Prioritization
  • Embedded AI
  • Extensible AI
  • Differential AI
  • Power Users
4

Responsible AI

Whether the AI is worth standing behind.

  • RAI Principles
  • RAI Measurement
  • RAI Mitigations
  • AI Governance
  • Oversight and Review
5

Scaling AI

The operations that let AI endure and grow.

  • AI Operations
  • Data Governance
  • Technical Debt
  • Monitoring and Metrics
  • Digital Fluency

Centru.ai

Where you put Centru to work.

Five modules, one model, with AI throughout: rating what you tell it, drafting what you’d otherwise write from scratch, and connecting your work back to the model. Microsoft ASSERT is inside, so every workload can be tested before a person signs it off.

Dashboard. Your AI practice on one screen: what needs attention now, every workload by release status, maturity by pillar, progress against each North Star, and what’s coming due.

Expertise

Ask the model

Ask about AI strategy, priorities, or a decision, such as which workload to fund next, how to make the case for an investment, or what to stop doing.

  • Answers grounded in Centru Core and IQ, with the reasoning on the page.
  • Bring your own material as text files, and pick up any conversation where you left off.

Maturity

Know where you stand, and what moves you forward

Assess your organization across the five pillars and twenty-five weighted dimensions.

  • Describe your AI program in your own words: Centru maps each answer to the dimensions it touches and rates it from Unaware to Strategic, as the evidence supports it.
  • Every rating explained, with findings and recommendations for your organization.
  • One shared conversation per organization, so a colleague can pick it up where you stopped. Or rate dimension by dimension yourself.
  • Reassess to see where you’ve improved, with a radar and heatmap that show leadership where to focus next.

Strategy

Strategy that gets executed

Set your North Star aspirations, the most important goals your organization is working toward, then the priorities that serve them and the planning horizons you’ll deliver across, and run the roadmap against all three.

  • Give every roadmap item an owner, a status, and a date, and see what’s on course, what’s at risk, and what’s coming due.
  • Work the roadmap as a board or a list.
  • Link every roadmap item to the AI workloads that deliver it.

Workloads

Take every AI workload from plan to launch

Keep the AI you’re buying and building in one place, and release it one stage at a time: pilot, limited availability, general availability.

  • For every release, record its scope, intended uses, and risks, chosen from a library of 48 risk types, from prompt injection to grounding errors, or named yourself.
  • Test each risk with Microsoft ASSERT. Your engineer’s coding agent connects to Centru over MCP and runs the evaluation in your own environment, and the results come back against the release they were run on.
  • Show the mitigations and safeguards in place, then send the release for review with the evidence attached. A person makes every release decision.
  • Centru’s AI can draft a workload or a release for you to review and refine.

Open specification

Centru Core is published in full, free for anyone to read.

Core grounds every assessment, observation, and recommendation Centru makes. Centru IQ and Centru Skills are built on it, and both grow continuously, with companion papers such as The Frontier Campus going deeper.

Plans

The full model on every plan.

Every subscription is an annual term with the full Centru 3 model, all five modules, and your choice of AI models. Plans differ in scale and in the support around them.

Available now

Centru Professional

For a single organization running its own AI strategy work.

  • 5 seats, with more available
  • One workspace
  • Self-guided onboarding
  • Standard support
Sign up
Available now

Centru Enterprise

For organizations running the model across business units, regions, or entities.

  • 15 seats, with more available
  • Multiple workspaces
  • A half-day onboarding session
  • A named support contact and 1:1 guidance from our experts
Talk to us
Licensed separately

Centru Engineering

Centru Skills enabled in your environment, with a forward deployed engineer alongside, for either plan.

  • Installed and configured in your environment by Cloud Lighthouse
  • Builds alongside the coding tools your teams already use
  • Can run fully offline where security demands it
About Forward Deployed Engineering
On request

Private cloud

For organizations that need Centru deployed in their own private cloud.

Talk to us

Preferred rates for non-profits, universities, UN system organizations, and Cloud Lighthouse clients. If you’re in a Counsel relationship or a Lab commitment, Centru.ai may already be included.

Where Centru fits

Agent 365 and Control Tower manage the AI you already have. Centru helps you decide what’s worth having, and keeps the audit trail from decision to release.

Complementary

Microsoft Agent 365

A control plane for the agents you already run: registry, access, and security, built for IT and security teams.

Agent 365 manages the agents you already run. Centru fits alongside it, helping you decide which are worth having, and keeping the audit trail from decision to release.

Complementary

ServiceNow Control Tower

Discovers built, bought, and shadow AI, and governs and measures it through the ServiceNow platform, with ROI and cost dashboards.

If you run ServiceNow, Control Tower can hold the operational inventory. Centru fits alongside it, helping you decide what belongs in it.

Centru is unique

Centru

Starts from a published, weighted model of what good looks like, runs diligence on you, holds strategy as a working instrument, and builds and tests in your own environment.

No platform prerequisite. It works in concert with Microsoft, Google, Amazon Web Services, Anthropic, and others.

The difference. Nothing else we’ve found combines self-diligence, strategy as a working instrument, workload evidence, plus building and evaluation in your own environment.

Get started

See where you stand.

Sign up and run your first AI Guided Assessment in Centru.ai, or start with an expert-led AI Maturity Assessment, benchmarked against the Centru AI Maturity Model.

About the AI Maturity Assessment