Training is the last step, not the first.
The request usually arrives as "our people need AI training." The need underneath is different: AI is already in the organization — in decisions, processes, tools that shipped with it switched on — and the ability to run it well is concentrated in a handful of people while the exposure is spread across everyone. That is not a course-catalogue problem. It is a capability problem, and it is organizational before it is individual.
- Written against
- EU AI Act · NIST AI RMF · ISO/IEC 42001 · ISO/IEC 42005 · ISO/IEC 23894 · ISO/IEC 38507 · ISO 31000 · ISO 21502
- Last updated
Literacy is the floor. Everything else stands on it.
Permalink to “Literacy is the floor. Everything else stands on it.”AI literacy is the ability of your people to work with AI knowingly. They can tell when a system is deciding and when it is assisting. They have a working sense of what it was trained to optimise. They know when an output deserves a second look, and where to take a concern. The EU AI Act asks providers and deployers to support this literacy in the people who operate their systems — and every capability on this page assumes it.
Above the floor there is a ladder. Some roles need more: the manager who owns a use case, the risk owner, the person who builds and runs the management system, the one who audits it. For them, literacy grows into qualification — and where it counts, certification.
One trained person is a bottleneck. The AI management system needs someone who can build it — and it needs the people around them able to work inside it: the manager who owns a use case and answers for its benefit, the risk owner who reads the exposure, the data lead who answers for what the models learned from, the people whose daily work the systems now touch. Send one person on an expert course and the organization acquires a translator. Build the set of roles, and it acquires a capability.
Which roles, at which depth, is not a matter of taste. The standards and regulations this work runs on imply who must be able to do what. Read them together and a structure appears: capabilities the organization needs, roles that carry them, and only then the training that builds each role.
Every organization that runs AI needs capability in the same four areas. What differs — and what decides what to build first — is your context: your systems, your obligations, your people.
- AI RiskFind, size, treat and watch AI risk — threat and opportunity alike. The roles that read exposure, and the qualifications that make their reading credible.
- AI ValueDecide what is worth doing, deliver it, prove the return. The roles that own a benefit and answer for it.
- AI GovernanceThe management system: direction, accountability, and the apparatus that makes both stick. The roles that build it, run it, and audit it.
- AI LiteracyAI competence for everyone — a common floor. What every employee should recognise about the AI in their work, and when to escalate.
A capability map is developed with you, not handed to you. In a short, fixed-scope engagement we map what running AI well requires of your organization: which capabilities matter in your context, which roles carry them, who holds them today, and where the gaps are.
The result is your map, not a template. The roles are named in your language. Each carries the qualifications that would support it — drawn from any relevant provider, certification where it earns its place, literacy measures where a course would be the wrong instrument. And it ends in a sequenced plan: who first, what can wait, what you already have and only need to recognise.
Teams that start from a catalogue buy what looks relevant. Teams that start from their map buy what their roles are missing — which is usually less, and always better placed.
Start with the map.
Tell us where AI touches your organization and who runs it today. We scope your capability map in one conversation — fixed scope, fixed price, delivered in weeks, with the qualification plan attached.