Workshop contents
Session 1 · Agentic AI
Agentic AI
Running without being asked
Until now, every action required someone to type a prompt. In this workshop, agentic AI means AI that can run an assigned task on a schedule or trigger, then report what it found.
Operating Loop
| Loop stage | What AI checks | Control question |
|---|---|---|
| Trigger | The agent runs at a planned time or when a condition is met. | Who approved this schedule? |
| Check | It reads the latest data and compares results against targets. | Is the data complete and current? |
| Alert | It flags problems and explains the evidence. | Who receives the alert? |
| Recommend | It drafts a response within defined limits. | What is it allowed to recommend? |
| Log | It records what it checked, found and sent. | Can the decision be reviewed later? |
Daily monitoring scenario
Name: Supply Chain Monitor Trigger: Runs daily at 6:00 AM INSTRUCTIONS: 1. Pull yesterday's delivery data 2. Calculate late delivery rate by region and shipping mode 3. Compare to the 30% target late-delivery rate 4. IF any region exceeds 50% late delivery rate: - Analyze contributing factors (delivery provider, product mix, order spike) - Generate alert with supporting evidence - Draft recommended action - Send summary to VP Logistics 5. IF all regions are below 30%: - Log "All clear" and update weekly trend tracker 6. Every Friday: generate weekly performance summary with trend analysis
What the output shows. On a bad day, the agent sends an alert with evidence and a recommended action. On a good day, it logs an all-clear. In the historical data, most regions sit between 53% and 58% late delivery, so the alert reflects current reality and the all-clear reflects the target state.
This agent runs every morning without being asked. But it depends on clean data, reliable execution, and clear rules for unusual cases. Who monitors the monitor?
Outcome. Agentic AI keeps working over time without being prompted. The executive question is not how many agents you deploy, but what each one may access, recommend, and execute.
Beyond a single agent
Organizations rarely run one agent. They run a small set of specialized agents that all report into one central summary. They should not freely negotiate with each other out of sight.
| Specialized roles | Each agent handles one distinct task: demand forecasting, delivery monitoring, delivery provider optimization, cost control. |
| Centralized output | The agents do not message each other. Each sends findings and alerts into one executive summary, so leaders see one connected view. |
| Governing boundaries | Every agent has limits. The defining question for each is what it may access, decide, and escalate to a person. |
The structure is what keeps this manageable. Several agents contribute to one unified picture for leadership, without hidden decision-making between the agents themselves.
If four agents disagree, who reconciles them, and does that person see the disagreement, or only the summary?