Role-specific learning
Learn the task. Practise the judgement.
Our learning approach starts with roles, real tasks and the consequences of getting them wrong. Participants practise checking sources, handling information, recognising limits and knowing when to stop or escalate.
What should role-based AI literacy enable people to do?
Role-based AI literacy should help people make sound decisions about the AI systems they use, build or oversee. A user needs to check an answer and protect information; a maker needs to understand access and release; a leader needs to judge benefit and consequences. Amplified Pi translates those responsibilities into practical learning and follow-up. Attendance records alone do not demonstrate competence or establish legal compliance.
- When this is a good fit
- The organisation is introducing AI across different roles and needs a shared baseline with practical depth appropriate to each responsibility.
- Before you invest
- Identify the actual systems, permitted uses and decisions people make. Agree learning objectives and proportionate records before choosing a generic course or promising certification.
Shared foundations. Different responsibilities.
Everyday users
Complete tasks and check results.
Champions
Support colleagues and make sense of questions.
Citizen developers
Build and hand over within clear rules.
Super users
Delegate multi-step tasks with control.
Platform and security owners
Manage controls, operation and incidents.
Sponsors and managers
Assess value, risks and investment.
Learning reflects the role, system and context of use. It supports AI-literacy work; a course alone does not establish compliance with the EU AI Act.
Everyday users
Complete a real task and check the result.
Source verification, confidentiality, hallucinations, bias, permitted tools and escalation.
In practice
A reviewed work sample and a personal task checklist.
Champions
Help colleagues without inventing policy.
Coaching, validated examples, inclusive support, feedback triage and clear escalation.
In practice
A facilitated clinic and a reviewed example for the community.
Citizen developers
Build within a defined operating boundary.
Data and connector policies, permissions, testing, change control, ownership and support.
In practice
A solution review with acceptance criteria and handover.
Super users
Delegate multi-step work with judgement.
Task scope, permissions, plans, meaningful approvals, prompt injection, intervention and recovery.
In practice
A bounded Cowork task with a review of actions and exceptions.
Platform and security owners
Operate and challenge the controls.
Identity, information protection, logs, evaluations, incidents, cost and feature changes.
In practice
A control exercise and an operational response scenario.
Sponsors and managers
Decide where to invest and where to stop.
Use-case value, risk ownership, workforce implications, quality evidence and limits of attribution.
In practice
A scale, adjust or stop decision against agreed criteria.
AI literacy in context
Our learning approach supports your AI-literacy work under the EU AI Act. Content and practice reflect roles, the systems in use and the consequences of their use.
Practical exercises and internal learning records are our proposed approach. A course alone does not establish compliance. Intended use and additional obligations require separate assessment.
Reference: European Commission guidance
Support learning and refresh it when work changes.
We use approved or synthetic materials, review work samples and provide everyday task aids. Content, participation and follow-up are recorded proportionately. Material changes to tools, roles or risks trigger a review.
The purpose, access and retention of learning records are agreed. Privacy specialists and employee representatives are involved where required before individual analytics are introduced. We do not create employee league tables.
How we approach the work
AI literacy becomes useful when people can make better decisions in the systems they actually use. We build a role-based learning programme around realistic work, observable judgement and support after the session.
-
Map roles to decisions and consequences
We identify the systems in use, who supplies information, who builds solutions and who approves or relies on results. Different responsibilities lead to different learning needs. The programme starts with a role and task inventory rather than the same presentation for everyone.
What you receive
A learning-needs map and agreed objectives for each audience.
-
Teach a shared baseline, then specialise
Everyone practises recognising limitations, protecting information, checking results and escalating uncertainty. Users then work on daily tasks; makers on permissions and release; platform teams on controls and incidents; leaders on investment and risk decisions. Examples use approved or synthetic information.
What you receive
A curriculum, exercises and task aids matched to actual responsibilities.
-
Observe judgement through realistic exercises
A participant might need to spot an unsupported statement, refuse an inappropriate data upload or recognise that a workflow exceeds their authority. We examine the reasoning and the corrected action. Attendance records alone do not show whether someone can handle the task.
What you receive
Reviewed exercises and a clear plan for additional practice.
-
Support transfer into everyday work
Managers provide time to apply the learning and champions help with recurring questions. We record content versions, participation and follow-up proportionately. Changes to tools, responsibilities or observed failures trigger a refresh. Records have an agreed purpose, access model and retention period.
What you receive
A follow-up routine and a maintained learning record.
The guidance behind the approach
Practical competence in its policy context
The European Commission’s AI-literacy material provides the policy context. Its own internal programme combines role-specific learning and peer exchange. Our proposed curriculum uses those ideas without presenting course completion as a legal compliance certificate.
Reference: European Commission: AI talent, skills and literacy; European Commission: its internal AI literacy programme
Illustrative example
Example: three roles, one task
A user checks a draft against its sources. A maker checks whether the workflow gives the right people access. A manager decides whether the result is suitable for the business decision. The learning programme uses the same scenario to practise those different responsibilities, instead of assuming every role needs the same technical depth.
What we need to get started
- Current tools, permitted uses and relevant internal policies
- Representatives from business, learning and platform teams
- Approved practice material and time to apply the learning
Questions before you begin
Do you provide an AI Act compliance certificate?
We can document participation and the agreed learning activities. That is not a certification of an organisation or a system. Applicable obligations depend on the intended use and current law, and require a separate assessment.
Is a one-off course enough?
Our proposed programme includes practice, support and refresh triggers. A course may establish knowledge, but the organisation still needs to help people apply it and respond when tools or responsibilities change.
Related modules
Next step