Trellis · Compliance Documentation
Framework alignment · Generated 9 November 2026
Enforcement begins August 2026
Article 4 of the EU AI Act (in force February 2025, enforced August 2026) requires all organisations deploying or using AI systems to ensure their staff possess a "sufficient level of AI literacy" — including understanding of AI capabilities, limitations, and the risks of automated decision-making.
trellis assessment records constitute documentary evidence of Article 4 compliance measures. Specifically, the following data serves as compliance evidence:
| Evidence type | Article 4 requirement addressed | Where in trellis |
|---|---|---|
| Assessment completion records | Demonstrates that AI literacy was measured for named staff | Manager dashboard → completion list |
| Stewardship pillar scores | Measures staff understanding of AI ethics, data privacy, and responsible use obligations | Team report → Five Pillar averages |
| Judgment pillar scores | Measures ability to critically evaluate AI outputs and detect errors or hallucinations | Team report → Five Pillar averages |
| Downloadable team report | Provides a timestamped, org-specific record of AI literacy measures taken | Manager dashboard → Download report |
Stewardship Pillar — Article 4 Alignment
The Stewardship pillar assessment (Instrument 2b) is conducted without AI assistance by design. It measures staff understanding of:
Practitioner-level Integrity scores (≥ 3.0) indicate staff are consistently applying responsible AI practices. Scores below 2.0 indicate a material compliance gap that should be addressed before the August 2026 enforcement deadline.
Five foundational content areas (February 2026)
The US DOL AI Literacy Framework defines five foundational content areas for AI literacy in the workplace. The Trellis Five Pillars map directly to these areas:
| DOL Content Area | Trellis Pillar | What is measured |
|---|---|---|
| FA1 — Understand AI Principles FA5 — Use AI Responsibly | Integrity | Ethics, data privacy, disclosure obligations, accountability for AI-assisted outputs |
| FA2 — Explore AI Uses | Scope | Task-AI fit, interaction mode selection, identifying where AI adds genuine value |
| FA3 — Direct AI Effectively | Craft | Prompt structure, context-setting, constraints, iterative refinement (CARE framework) |
| FA4 — Evaluate AI Outputs | Judgment | Critical evaluation, hallucination detection, domain-grounded verification |
| Continuous learning principles | Momentum | Adaptability as AI evolves, proactive use-case identification, peer knowledge sharing |
European Commission Digital Competence Framework (December 2025)
DigComp 3.0 is the EU's official digital competence framework, updated to integrate AI competence across all five competence areas. The Trellis Five Pillars are fully aligned:
| DigComp 3.0 Area | Trellis Pillar |
|---|---|
| CA1 — Information and Data Literacy | Judgment |
| CA2 — Communication and Collaboration | Momentum |
| CA3 — Digital Content Creation | Craft |
| CA4 — Problem-Solving with Digital Technologies | Scope |
| CA5 — Safety, Wellbeing & AI Ethics | Integrity |
Industry benchmark for structuring AI competency levels
Trellis uses the Gartner five-level maturity model for scoring. Each pillar is scored 1–5, corresponding to these maturity levels:
| Level | Label | Description |
|---|---|---|
| 1 | Emerging | Aware that this competency matters; no consistent practice yet. |
| 2 | Developing | Building foundational behaviours; applies them in familiar situations. |
| 3 | Practitioner | Applies the competency consistently and independently; can articulate reasoning. |
| 4 | Advanced | Extends to novel situations; develops tools or guidance for colleagues. |
| 5 | Leader | Shapes team practice; mentors others; contributes to standards and governance. |