From experiment to industrial scale

An AI transformation that holds rests on six pillars. Assess your organisation in six questions: you get your maturity level and the acceleration levers, role by role.

Assess my maturity See the six pillars

The six pillars of an AI transformation

None makes up for the others. A brilliant model on siloed data produces nothing; so does impeccable data with no adoption.

  • Strategy

    An arbitrated, budgeted roadmap tied to business objectives. Without it, AI scatters into isolated initiatives nobody consolidates, “shadow AI”.

  • Data

    The raw material. Reconciled sources, measured quality, explicit governance: this is the pillar that decides what the other five can produce.

  • Technology

    Infrastructure that ships to production and holds under load: continuous integration, monitoring, the ability to roll back without ceremony.

  • Talent

    The in-house skills that make the organisation self-sufficient. A contractor builds; a team keeps it alive.

  • Governance

    Risk control: confidentiality, bias, traceability of decisions. It is designed with the system, never after the incident.

  • Culture

    Real adoption, not the number of licences. It is the slowest pillar to move and the one that decides the return on investment.

Assess your maturity in six questions

One answer per pillar. Nothing is sent: the calculation happens in your browser.

  1. 01 · StrategyIs there a clear AI roadmap?
  2. 02 · DataWhat state is your data in?
  3. 03 · TechnologyDo you have production-grade infrastructure?
  4. 04 · TalentDo you have the skills in house?
  5. 05 · GovernanceHow do you handle risk: confidentiality, bias, traceability?
  6. 06 · CultureWhat is real adoption in the field?

The four maturity levels

The average of your six answers places the organisation. The goal is not to reach level 4 everywhere, but to close the gap between the highest and the lowest pillar.

  1. Initiation

    Level 1 · Exploratory

    Isolated initiatives, siloed data. The value produced stays local and leaves with the person who created it.

    The next step: reconcile two sources that matter and measure the quality you get.

  2. Structuring

    Level 2 · Practising

    The first wins are real and the team is learning. Going to production is still hand-made: every release has to be negotiated.

    The next step: industrialise a case already proven rather than opening a seventh.

  3. Acceleration

    Level 3 · Systemic

    AI is built into the processes and monitored like the rest of the system. Models live, get replaced and are traced.

    The next step: open the capability to the business without going through IT for every request.

  4. AI-first

    Level 4 · Transforming

    AI sits at the core of the business model. Decisions rest on real-time data whose chain is auditable end to end.

    The next step: hold the level, because governance debt is the first risk at this stage.

The levers, role by role

What goes into production at each level, without skipping a step.

Role Level 1 Level 2 Level 3 Level 4
Information systems Automate level 1 tickets Automate employee onboarding Predictive maintenance Automatic incident remediation
Human resources Go paperless on files Assisted self-service Strategic workforce planning Assisted career guidance
Finance Remove invoice re-keying Automatic order, receipt and invoice matching Continuous budget forecasting Real-time fraud detection
Operations A single digital inventory Real-time flow tracking Automatic replenishment A complete digital twin

Each row assumes the previous one: predictive maintenance is only worth it once the inventory is already reliable.

Your profile, then a plan

Take the assessment, send us the profile you get: we come back with the two projects that move your score the most.

Take the assessment See the plans