AI Systems
First 2027 scholarship call
AI Agents
Design agentic systems that reason over goals, use tools, retain context and escalate safely.
Developed under the new AI Act model · Continuous updates
Edition 2026
200 hours of applied learning
8 academic credits
25 lessons + 6 workshops
Digital campus
Verifiable certificate
550 USD one payment · or 2 × 275 USD (today and after 30 days)
What you'll learn
- Justify an agent or refuse it.
- Bind every tool to a permission.
- Keep memory apart from instructions.
- Put a human on money.
- Read an action log.
- Name the stop and its owner.
- Apply the IntellCore AI Act Method™: classify the risk of an AI use case in your field and document it under the AI Act 2026.
Program curriculum
01 — What an agent is (and is not)
- What AI is and how it reached agents
- A chat is not an agent
- The four-step loop
- Levels of agency (least agency)
- The case of the week and the sheet
02 — Anatomy: six objects
- The six pieces
- Memory is not pasting the manual
- Non-human identity
- Guardrails you can audit
03 — One agent or many
- The myth of the agent team
- The token trap and the deadlock
- Topologies that actually exist
- A written decision
04 — Tools, MCP and computer use
- From reading to acting
- MCP, the pipe
- The lethal trifecta
- Designing the case's tools
05 — Risk, OWASP and who signs
- Governance is missing
- ASI01 to ASI10, in plain words
- Five questions and the cut-off
- Owner
06 — Agent pilot dossier
- What an agent dossier is
- Eight slides
- The table
- Switch on, measure, switch off
07 — IntellCore AI Act Method™ applied to your field
- The EU AI Act 2026: risk levels, prohibited practices and obligations
- Transparency, human oversight and documentation in your professional field
- Classifying and auditing a real AI use case step by step
- Your compliance file with the IntellCore AI Act Method™: from system records to the review plan
Practical application
- · Agent pilot dossier: a level-1 or narrow level-2 refund agent, with a one-page sheet, allow-listed tools, logging, an owner and a stop.
Practical project: Agent pilot dossier: a level-1 or narrow level-2 refund agent, with a one-page sheet, allow-listed tools, logging, an owner and a stop.
Who is this for?
- · Whoever would connect a model to a business system
- · Operations, customer service and IT leads
- · Teams assessing "agentic" products
Requirements
- · No coding required
- · About 6 hours per week
Meet your instructor
AI Agents & Automation
Chair, AI Agents & Automation
Tool calling, planning loops, state handling, guardrails and human-in-the-loop checkpoints.
Teaches: AI Automation
Edition 2026
What each week includes
Every week of the program follows the same structure.
- 01
Explainer video per lesson
- 02
Guided reading (6 h)
- 03
IntellCore Labs · ~90 min workshop with data
- 04
Verified real case
- 05
IntellCore Podcast
- 06
IntellCore Radar
- 07
Self-assessment
- 08
Practice and exam (70%, 3 attempts)
- 09
Extended practice (5 h)
- 10
IntellCore AI Act Method™ module applied to your field (30 h)
- 11
Applied final project (30 h)
- 12
6 modules × 23 h + IntellCore AI Act Method™ module 30 h + final assessment 2 h + final project 30 h = 200 h
- 13
Downloadable PDF summary
Verifiable certificate
200 hours · 8 academic credits
Unique verifiable code at intellcore.ai/verify
Student reviews
Reviews are published from verified students only, after the first cohort completes the program. We do not publish testimonials we cannot attribute.
