What 40+ real AI literacy initiatives reveal about how organisations train for AI.

The EU AI Office’s Living Repository of AI literacy practices is a public catalogue of how organisations train people for AI — not a ranking, not a certificate list, and not a vendor shortlist. FluencyRank read the current official inventory (42 fact pages at the August 2026 scan) for recurring patterns. This essay reports those patterns. It does not claim EU endorsement of FluencyRank, and matching a listed practice does not prove Article 4 compliance.

Distinguish common patterns from isolated examples. Many mature rows describe a baseline for all staff, then role-specific depth — for example legal or public-affairs tracks beside generalist paths. Governance context shows up often: approved tools, policy reminders, risk-aware guidance. Champions and communities of practice appear across sectors. Short, repeated learning shows up more than a single annual workshop. Measurement, where stated, is usually participation, hours, or reach — not independently validated competence.

Inclusion in the Living Repository is not Article 4 compliance, not a Commission evaluation of FluencyRank, and not permission to sell a “repository-aligned” certificate.

Patterns that keep recurring

• Baseline for everyone, then role-specific depth — not one generic LMS module for all jobs
• Literacy tied to policy and approved tools — practice sits next to governance, not instead of it
• Champions and peer networks — ambassadors and communities, not only top-down slides
• Activity measured more often than demonstrated judgment — attendance is not capability
• Continuous refresh is uneven — some programmes update; many fact pages leave refresh unstated
• Use-context awareness varies — named systems and workplace cases in some rows; generic awareness in others

Those are aggregated observations from official Commission repository and listing pages. When we discuss a specific organisation in sales or research notes, we link the individual fact page. Here the point is the pattern: organisations treat literacy as an operating mix — baseline, role depth, governance — not a one-off course.

Where FluencyRank aligns — and where it intentionally differs

• Aligns — daily workplace practice, role paths, policy-in-scenario reminders, champion-facing evidence of measures
• Aligns — separates factual practice evidence from derived readiness views; no universal literacy certificate
• Differs — FluencyRank is employee practice inside a paying organisation, not citizen education or a public academy
• Differs — no belts, ECTS, coding bootcamps, or runtime agent product; no Art. 4 safe-harbour export

The Commission’s AI literacy Q&A reinforces the same humility: literacy measures should fit role and context. FluencyRank uses that as a design constraint — short role-relevant challenges and evidence champions can show — not as a claim that completing packs satisfies regulatory duties. Your counsel and compliance function still own conclusions.

Readiness views reflect platform engagement. They are decision support for L&D and champions, not a comprehensive capability assessment or a compliance attestation.

If you want to see how practice evidence looks without mistaking it for a repository badge, start with a sample readiness report — then compare your own programme to the patterns above, using the Commission catalogue as the canonical source.