The same four levels, taught through the problems each discipline actually faces. Universities pick the tracks that fit their faculties. Students can cross over once they reach Practitioner.
AI-assisted codingPair programming that you still understand
Models as componentsAPIs, routing, cost and quality trade-offs
Agents that shipTools, memory, approvals, testing, deployment
Production AIEvaluation at scale, sovereignty, a capstone
AI for defendersTriage, log reading, threat summaries
Securing AI systemsPrompt injection, data leakage, the agent firewall
Red-teaming AIAttacking agents before release, rating them
AI security architectureTrust layers, audit, a capstone
Asking data questionsNatural language to analysis
Pipelines with AICleaning, enrichment, classification at scale
Analytical agentsReports that write and check themselves
Decision systemsForecasting, governance, a capstone
Document assistants, compliance agents that log their own reasoning, policies a regulator would accept.
Safe use in care, clinical documentation that cites and defers, care-pathway agents with a human in charge.
Content with judgement, campaign assistants, customer agents that stay on brand and on policy.
Each track starts with what is allowed and what is not, moves to assistants for the work your team repeats, and ends with agents that keep a human in charge.
| Track | Novice | Practitioner | Builder | Expert |
|---|---|---|---|---|
| Business & management | AI at workResearch, writing, meetings, decisions | Process assistantsAutomating the work your team repeats | Operations agentsApprovals, exceptions, cost control | Leading AI adoptionBusiness cases, risk, a capstone |
| Marketing & communications | Content with judgementDrafting, tone, brand, disclosure | Campaign assistantsSegments, variants, measurement | Customer agentsConversations that stay on brand and on policy | Growth systemsAttribution, experiments, a capstone |
| Law & governance | AI and the lawWhat is allowed, what is not, what is unclear | Document assistantsReview, clause comparison, summaries with sources | Compliance agentsChecks that log their own reasoning | AI governancePolicies a regulator would accept, a capstone |
| Finance & accounting | AI for numbersReconciliation, variance, explanation | Close assistantsMatching, exceptions, journal preparation | Finance agentsApprovals, controls, audit trail | AI in audit and riskAssurance over AI, a capstone |
| Design & media | Generative toolsImages, video, voice, and their limits | Production workflowsBriefs to drafts to approvals | Creative agentsPipelines that keep a style and a licence | AI-native studiosRights, provenance, a capstone |
| Health & life sciences | Safe use in careSummaries, literature, what never to trust | Clinical documentationAssistants that cite and defer | Care-pathway agentsScheduling, follow-ups, with a human in charge | Regulated AIValidation, privacy, a capstone |
| Education | Teaching with AILesson design, feedback, integrity | Tutor assistantsMarking support, differentiated materials | Learning agentsAdaptive practice with teacher control | AI in the institutionPolicy, assessment, a capstone |
| Track | Novice | Practitioner | Builder | Expert |
|---|---|---|---|---|
| DeveloperComputing, software engineering | AI-assisted codingPair programming that you still understand | Models as componentsAPIs, routing, cost and quality trade-offs | Agents that shipTools, memory, approvals, testing, deployment | Production AIEvaluation at scale, sovereignty, a capstone |
| CybersecurityInfocomm security, networks | AI for defendersTriage, log reading, threat summaries | Securing AI systemsPrompt injection, data leakage, the agent firewall | Red-teaming AIAttacking agents before release, rating them | AI security architectureTrust layers, audit, a capstone |
| Data & analyticsData science, statistics, IS | Asking data questionsNatural language to analysis | Pipelines with AICleaning, enrichment, classification at scale | Analytical agentsReports that write and check themselves | Decision systemsForecasting, governance, a capstone |
Start with one faculty and one track. Add the rest once your first cohort has a portfolio to show.
