# AI Systems Expert > An online, self-paced AI course from ARFA, the TIBLOGICS AI Academy. Advanced level, about 21.5 hours, in English and French, with a verifiable certificate. Web page: https://tiblogics.com/learning-box/ai-systems-expert ยท All tracks: https://tiblogics.com/learning-box.md ## Key facts - Level: Advanced - Time: about 21.5 hours including hands-on work (6 modules, 25 lessons, 6 labs); self-paced, no deadline - Price: $677 one time for lifetime access to this track, or every track for $89 a month (cancel anytime) - Certificate: TIBLOGICS Certified AI Systems Expert, with a public verification page - Languages: English and French ## About This is the level for people who will decide how AI is used, not only use it. It opens with systems thinking at organisational scale: stocks and flows, delays, reinforcing and balancing loops, incentives, and where to intervene. Every module after that applies it. You will learn what agents really are and when a simpler workflow is the better choice, how to evaluate and monitor AI systems, how they fail and how to secure them, how risk-based regulation and good documentation work, and how to cost, measure and lead adoption honestly. The labs are real design work done inside the platform, and the capstone is a full AI system and leadership plan for an organisation you know, reviewed by a person. Skills this track builds also appear in generative AI leadership and AI governance certifications. This track is independent: it is not affiliated with any vendor or certification body and is not official exam preparation. Who it is for: Practitioners who use AI confidently and now lead it: team leads, heads of operations, product and technology leaders, consultants and founders. ## What you will be able to do - Model an AI rollout as a system of loops, delays and incentives, and find the leverage point - Decide between a fixed workflow and an agent, and scope an agent's permissions safely - Design an evaluation set and monitoring that catch problems before users do - Defend against prompt injection and plan incident response - Assess risk, document accountability and run vendor due diligence - Cost AI honestly, measure value against a baseline, and lead adoption ## Curriculum ### Module 1: Systems Thinking for AI at Scale See an organisation's AI use as a system of stocks, flows, loops and incentives, so you intervene where behaviour actually changes rather than where it is easiest to push. - Stocks, flows and delays (30 min) - Reinforcing and balancing loops in AI adoption (30 min) - Metrics, incentives and Goodhart's law (30 min) - Leverage points: where to intervene (30 min) - Module quiz ### Module 2: Agents and Tool Use Understand agents as goal-pursuing loops, give them only the access they need, and choose a simpler fixed workflow whenever it will do the job. - What an agent actually is (25 min) - Tools, permissions and least privilege (30 min) - Workflows versus agents: choose the simpler thing (25 min) - When agents go off-script (30 min) - Module quiz ### Module 3: Evaluation and Monitoring Build the evaluation sets, graders and production monitoring that tell you, before your users do, whether an AI system is still doing its job. - Designing an evaluation set (30 min) - Graders and their failure modes (30 min) - Regression testing when models change (25 min) - Monitoring in production (30 min) - Module quiz ### Module 4: Security and Failure Modes Treat AI security as a system of inputs, permissions and exits, then design guardrails, incident response and red-teaming that hold up when the model does not. - Prompt injection and data exfiltration (30 min) - Guardrails that work: allow-lists, approvals, logs (30 min) - Incident response for AI systems (30 min) - Red-teaming your own system (30 min) - The 30 doors: a launch gate for AI systems you commission (30 min) - Module quiz ### Module 5: Governance, Risk and Regulation Build governance as a working system of rules, owners and feedback: understand risk-based regulation, document decisions, vet vendors and write a policy people follow. - Risk-based regulation and the EU AI Act (35 min) - Documentation and accountability (30 min) - Vendor and model due diligence (30 min) - An AI policy people actually follow (25 min) - Module quiz ### Module 6: Leading AI Adoption Lead AI adoption as a system: cost it honestly, measure value against baselines, design for the loops of trust, skill and incentive, and build a roadmap with kill criteria. - Costing AI honestly (35 min) - Measuring value with baselines, not anecdotes (30 min) - Adoption is a system (30 min) - Building a roadmap (30 min) - Module quiz ## How you are assessed - A quiz after each module (80% to pass, retakes allowed). - A timed final exam: 35 questions in 75 minutes, 75% to pass, 90% for distinction, 3 attempts. - A capstone project reviewed by a person against a published rubric (70% to pass). - Each certificate has a public page at https://tiblogics.com/certificates/ that anyone can open to check it.