AXIS 01/05 · DATA Where does the data that would feed an AI model live today?
L0 In spreadsheets and email threads scattered across teams.
L1 In several systems that do not talk to each other.
L2 Centralized in one database, with partial integrations.
L3 In a governed, documented data warehouse or lakehouse.
AXIS 01/05 · DATA If leadership asked tomorrow for a report crossing three business areas, what happens?
L0 Weeks of manual work and conflicting versions.
L1 Someone builds it by hand in days; the result depends on who does it.
L2 Dashboards exist, but adding new sources takes time.
L3 Minutes: defined metrics and a single source of truth.
AXIS 02/05 · PROCESSES How well documented are the processes you would want to automate?
L0 They live in people’s heads.
L1 There are manuals, but they are outdated.
L2 Key processes are mapped, with clear owners.
L3 Mapped, measured, with cycle metrics per step.
AXIS 02/05 · PROCESSES How much of an analyst’s day goes to capturing, copying or reconciling information between systems?
L0 Most of it: the operation is essentially manual.
L1 A lot: there are islands of automation, but people are the glue.
L2 Little: critical flows are automated; manual exceptions remain.
L3 Minimal: end-to-end orchestrated flows, with alerts.
AXIS 03/05 · TALENT Who would drive an AI project inside the organization?
L0 No one assigned: it would be extra work for somebody.
L1 An enthusiast, with no protected time or mandate.
L2 An owner with a mandate, supported by IT.
L3 A team with a product owner, data skills and executive sponsorship.
AXIS 03/05 · TALENT What real experience does the team have with AI tools?
L0 Almost none; there is skepticism or fear.
L1 Informal personal use (chatbots), with no guidelines.
L2 Isolated pilots in some areas, with mixed results.
L3 Daily use with measured cases and ongoing training.
AXIS 04/05 · GOVERNANCE If a model touched personal or sensitive data, what framework governs it today?
L0 None, explicitly.
L1 General IT policies, with no mention of AI.
L2 Privacy and access guidelines defined; AI policy still a draft.
L3 An active AI-use policy: data classification, roles and audit trail.
AXIS 04/05 · GOVERNANCE Who decides — and how — whether an AI-generated answer is acceptable for production?
L0 It has not been discussed.
L1 Each user’s individual judgment.
L2 Defined human review for critical cases.
L3 Quality criteria, evaluations and traceability per use case.
AXIS 05/05 · INFRASTRUCTURE Where does your technology run today?
L0 Aging on-premise servers and desktop software.
L1 Basic hosting or an in-house datacenter without redundancy.
L2 Partial cloud: some workloads migrated, others legacy.
L3 Managed cloud with IaC, monitoring and autoscaling.
AXIS 05/05 · INFRASTRUCTURE How easy is it to expose a new service (an API) on top of your current systems?
L0 Very hard: closed systems or no vendor to call.
L1 Possible, but every integration is a long project.
L2 APIs exist for the core; standardization is missing.
L3 Service architecture with documented APIs and CI/CD.
