AI Leaders Track
AI Agents & Intelligent Automation
Implementation-focused training for AI leaders who must design, govern and scale agentic systems. Understand how agents reason, plan, use tools and interact with your stack so automation stays trusted and under control.
Programme overview
AI Agents & Intelligent Automation is an advanced, implementation-first programme for AI leaders and senior practitioners who architect, deploy and govern agentic systems. It focuses on how agents really work, fail and scale—no generic GenAI evangelisation.
Who this programme is for
Target audience
- Head of AI / AI Lead
- AI or ML Architect
- Staff or Principal Engineer
- AI Platform Lead
- Technical CTO with hands-on system responsibility
Not designed for: business leaders, strategy-only roles, innovation teams without delivery ownership, or non-technical decision makers.
Prerequisites
- Solid understanding of LLM workflows (prompting, RAG, fine-tuning concepts).
- Experience with APIs, distributed systems or ML pipelines.
- Comfort with production constraints such as latency, cost, reliability and security.
Learning objectives
Agentic AI foundations
- Differentiating LLM workflows, automation pipelines and agentic systems.
- Understanding when a system should or should not be agentic.
- Mastering core building blocks: reasoning, planning, memory, tool use and feedback loops.
Architecture & implementation
- Designing end-to-end agent architectures tailored to real environments.
- Implementing single-agent and multi-agent patterns.
- Orchestrating agents with tools, APIs and data layers.
- Choosing control models (centralised, decentralised or hybrid).
Governance, control & reliability
- Defining autonomy boundaries and permissions.
- Implementing guardrails with human-in-the-loop/on-the-loop mechanisms.
- Ensuring traceability, observability and auditability of agent behaviour.
Detailed programme
1. Agentic AI fundamentals
- What “agentic AI” actually means (and what it does not).
- Agents vs chatbots vs workflow automation.
- Cognitive loop models: perception → reasoning → action → feedback.
- Why most “AI agents” fail in production.
2. Modern agent architectures
- Single-agent vs multi-agent systems.
- Tool-using agents (APIs, code execution, retrieval).
- Planning patterns: ReAct, Plan-and-Execute, Reflexion-based loops.
- Event-driven and stateful agent architectures.
3. Multi-agent systems & coordination
- Specialised agent roles and responsibilities.
- Inter-agent communication, delegation and arbitration.
- Conflict handling, consensus mechanisms and orchestration vs swarm approaches.
4. Intelligent automation vs agentic automation
- Identifying when an agent is the wrong solution.
- Hybrid architectures (RPA + LLM agents).
- Decision automation vs execution automation.
- Spotting over-engineering risks.
Analytical focus: deciding when not to build an agent.
5. Governance, safety & control
- Limiting agent action space, sandboxing and permission models.
- Monitoring reasoning, decisions and behaviour.
- Logging, replayability, audits and red-teaming agentic systems.
6. From prototype to production
- Assessment & certification and testing strategies.
- Cost, latency and scalability trade-offs.
- Observability metrics that matter.
- Managing long-term “agentic debt”.
Teaching approach
- Architecture-level design sessions.
- Deconstruction of real agentic systems.
- Pattern-based analysis of what works vs what breaks.
- Technical workshops focused on system design.
- Expert-level guided discussions (no marketing demos, no low-value live coding).
Key takeaways & deliverables
- Agentic AI architecture framework.
- Governance and safety checklists.
- Decision matrices: agent vs automation.
- Production-ready design patterns and reference architecture.
Positioning statement
Designed for organisations already serious about AI who now need agentic systems to be engineered, governed and controlled—not merely demonstrated. Complements executive AI strategy programmes without overlapping them.
Assessment & certification
Each participant validates outcomes through practical deliverables and receives documented feedback linked to certification requirements.
- Capstone pitch or technical runbook evaluated by faculty.
- Rubrics aligned with the programme’s KPIs and maturity model.
- Digital certificate of completion plus guidance on next credentials.
- Optional SOC or board readout to anchor sponsorship.