Phase 4: AI for Business and IT (Weeks 13–20)
Objective: This is the most important phase of the program. This is where prior experience merges with AI to create a high-value professional profile. Real AI projects are built to demonstrate business impact.
🎯 Expected Outcomes
By completing this phase:
- Deep understanding of Generative AI and its applications
- Advanced prompt engineering for enterprise scenarios
- Knowledge of AI Governance (NIST AI RMF, ISO 42001)
- Applied Responsible AI principles
- Microsoft Copilot and agents mastered
- 2 complete AI projects in the portfolio
- Portfolio website published
Week 13: Generative AI Foundations
Objective
Understand how LLMs work, their capabilities, and limitations.
Main Resources
| Resource | Language | Type | Duration |
|---|---|---|---|
| DeepLearning.AI: Generative AI for Everyone | 🇬🇧 English (ES subs) | Course | 5 hrs |
| Microsoft Learn: Generative AI | 🇪🇸 Spanish | Learning Path | 4 hrs |
| Google: Introduction to Generative AI | 🇬🇧 English | Course | 3 hrs |
| Coursera: Generative AI Concepts | 🇬🇧 English (ES subs) | Course | 4 hrs |
| IBM: Generative AI Fundamentals | 🇬🇧 English | Course | 3 hrs |
Daily Plan
| Day | Topic | Resource |
|---|---|---|
| Monday | What are LLMs? Transformers, tokens | DeepLearning.AI |
| Tuesday | Capabilities: text generation, summarization, code | Microsoft Learn |
| Wednesday | Limitations: hallucinations, bias, knowledge cutoff | DeepLearning.AI |
| Thursday | Enterprise use cases | Google course |
| Friday | Evaluation: when to use AI vs. not? | Reflection + journal |
Key Concepts
LLM Architecture:
├── Pre-training (massive data)
├── Fine-tuning (specialized tasks)
├── RLHF (human feedback alignment)
├── Inference (generating responses)
└── Limitations
├── Hallucinations
├── Knowledge cutoff
├── Context window limits
└── Bias amplification
Week 14: Prompt Engineering
Objective
Master the art of communicating effectively with AI models to obtain professional results.
Main Resources
| Resource | Language | Type | Duration |
|---|---|---|---|
| DeepLearning.AI: ChatGPT Prompt Engineering | 🇬🇧 English | Course | 2 hrs |
| Microsoft: Prompt Engineering Techniques | 🇪🇸 Spanish | Docs | 3 hrs |
| OpenAI Prompt Engineering Guide | 🇬🇧 English | Guide | 2 hrs |
| Anthropic Prompt Engineering | 🇬🇧 English | Guide | 2 hrs |
| Google: Prompt Design | 🇬🇧 English | Guide | 2 hrs |
Daily Plan
| Day | Technique | Practice |
|---|---|---|
| Monday | Zero-shot, few-shot prompting | 10 practice prompts |
| Tuesday | Chain of Thought (CoT) | Step-by-step analysis |
| Wednesday | Role prompting + system prompts | Create specialized personas |
| Thursday | Output formatting (JSON, tables, reports) | Enterprise templates |
| Friday | Prompt chains and workflows | Multi-step analysis |
🔨 Project: Professional Prompt Library
Create a documented collection of prompts for:
Security:
- Security incident analysis
- IOC generation from logs
- Executive vulnerability summaries
Compliance:
- Policy review against frameworks
- Automated gap analysis
- Compliance checklist
Management:
- Executive project summaries
- Risk analysis
- Action plans
Deliverable: GitHub repository with documented, categorized prompts and input/output examples.
Week 15: AI Governance
Objective
Understand the AI governance frameworks organizations are adopting.
Main Resources
| Resource | Language | Type | Duration |
|---|---|---|---|
| NIST AI Risk Management Framework | 🇬🇧 English | Framework | 6 hrs |
| ISO/IEC 42001:2023 Overview | 🇬🇧 English | Standard | 3 hrs |
| EU AI Act Summary | 🇬🇧 English | Regulation | 4 hrs |
| OECD AI Principles | 🇪🇸 Spanish available | Principles | 2 hrs |
| Microsoft AI Governance Framework | 🇬🇧 English | Blueprint | 2 hrs |
Daily Plan
| Day | Framework | Focus |
|---|---|---|
| Monday | NIST AI RMF: Govern function | Policies and roles |
| Tuesday | NIST AI RMF: Map function | Identify AI risks |
| Wednesday | NIST AI RMF: Measure function | Metrics and evaluation |
| Thursday | NIST AI RMF: Manage function | Response and monitoring |
| Friday | ISO 42001 + EU AI Act overview | Global regulatory context |
Key Concepts: NIST AI RMF
NIST AI Risk Management Framework:
├── GOVERN
│ ├── Policies & procedures
│ ├── Roles & responsibilities
│ └── Organizational culture
├── MAP
│ ├── Context & use cases
│ ├── Risk identification
│ └── Stakeholder engagement
├── MEASURE
│ ├── Metrics & benchmarks
│ ├── Testing & evaluation
│ └── Bias assessment
└── MANAGE
├── Risk prioritization
├── Response strategies
└── Continuous monitoring
Week 16: Responsible AI
Objective
Apply responsible AI principles in the design and deployment of systems.
Main Resources
| Resource | Language | Type | Duration |
|---|---|---|---|
| Microsoft Responsible AI Resources | 🇪🇸 Spanish | Framework | 4 hrs |
| Microsoft Learn: Responsible AI | 🇪🇸 Spanish | Learning Path | 4 hrs |
| Google: Responsible AI Practices | 🇬🇧 English | Guide | 3 hrs |
| IBM AI Ethics | 🇬🇧 English | Framework | 2 hrs |
| Partnership on AI | 🇬🇧 English | Resources | Reference |
Daily Plan
| Day | Principle | Practice |
|---|---|---|
| Monday | Fairness & inclusion | Evaluate model bias |
| Tuesday | Reliability & safety | Edge-case testing |
| Wednesday | Privacy & security | Data handling assessment |
| Thursday | Transparency & explainability | Model cards, documentation |
| Friday | Accountability | Governance workflow design |
🔨 Project: Responsible AI Impact Assessment
Create a complete assessment for a fictional AI system:
- System Description: Purpose, data, stakeholders
- Fairness Analysis: Impacted groups, potential bias
- Risk Evaluation: Severity × likelihood
- Mitigation Plan: Technical and process controls
- Monitoring Plan: Continuous metrics
- Transparency: User documentation
Deliverable: Complete document in GitHub with reusable template.
Week 17: Enterprise Copilot and Advanced Agents
Objective
Scale from basic agents (already built in Phase 1) to enterprise scenarios: autonomous agents, M365 Copilot integration, adoption, and governance.
Building your first agent in Copilot Studio moved to Phase 1 as a priority module. This week assumes you already have a published agent and focuses on advanced capabilities and enterprise strategy.
Main Resources
| Resource | Language | Type | Duration |
|---|---|---|---|
| Microsoft 365 Copilot Adoption | 🇬🇧 English | Adoption guide | 3 hrs |
| Learning Path: Prepare your organization for M365 Copilot | 🇪🇸 Spanish | Learning Path | 6 hrs |
| Applied Skill: Enhance agents with autonomous capabilities | 🇬🇧 English | Credential (lab) | 3-4 hrs |
| Copilot Studio Agent Academy (advanced ranks) | 🇬🇧 English | Practical curriculum | Variable |
Daily Plan
| Day | Topic | Practice |
|---|---|---|
| Monday | Autonomous agents: event-based triggers | Add autonomous behavior to your agent |
| Tuesday | Advanced agent flows + connectors | Integrate an external system (Dataverse/API) |
| Wednesday | M365 Copilot: extend with declarative agents | Agent over organization data |
| Thursday | Governance: DLP, authentication, RAI for agents | Agent security review |
| Friday | Adoption strategy and enterprise rollout | Deployment plan + value measurement |
Weekly goal: earn the "Enhance agents with autonomous capabilities" Applied Skill, which validates agent flows, autonomous triggers, and human hand-off.
Week 18: AI Project #1 — Security Incident Assistant
Objective
Build a functional AI assistant that demonstrates programming, AI, and security skills.
Project Specification
Security Incident Assistant:
- Receives a security incident description
- Classifies severity automatically
- Suggests response steps based on type
- Generates executive summary for management
- Recommends IOCs to search for
Technical Stack
Security Incident Assistant
├── Python (FastAPI o Streamlit)
├── OpenAI API / Azure OpenAI
├── Prompt engineering
│ ├── System prompt (security expert persona)
│ ├── Few-shot examples (incidentes clasificados)
│ └── Output format (JSON structured)
├── Knowledge base (playbooks de respuesta)
└── Documentation (README, architecture, usage)
Weekly Plan
| Day | Task |
|---|---|
| Monday | Design architecture, define prompts |
| Tuesday | Project setup, basic API calls |
| Wednesday | Implement classification + response |
| Thursday | Basic UI (Streamlit) + testing |
| Friday | Documentation + deploy to GitHub |
| Saturday | Polish + professional README |
Deliverable
Complete GitHub repository with:
- Functional code
- Professional README with screenshots
- Documented architecture
- Installation instructions
- Usage example
Week 19: AI Project #2 — Risk Assessment Assistant
Objective
Build a second AI project focused on enterprise risk analysis.
Project Specification
Risk Assessment Assistant:
- Receives information about a system or project
- Evaluates risks against a framework (NIST, ISO)
- Generates a prioritized risk matrix
- Proposes mitigation controls
- Exports a professional report
Technical Stack
Risk Assessment Assistant
├── Python
├── Azure OpenAI / OpenAI API
├── Prompt engineering
│ ├── Framework knowledge (NIST, ISO 27001, ISO 42001)
│ ├── Risk categorization
│ └── Control recommendations
├── Output formatting (Markdown report)
└── Export capability (PDF or HTML)
Deliverable
Professional repository with the same standard as Project #1.
Week 20: Portfolio Website
Objective
Create a professional website showing all completed work.
Resources
| Resource | Type | Use |
|---|---|---|
| GitHub Pages | Free hosting | Deploy |
| Hugo or Docusaurus | Static site generator | Framework |
| Tailwind CSS | Styling | Design |
Portfolio Content
- About Me — Professional bio, experience, vision
- Skills — Technical competencies with evidence
- Certifications — AZ-900, SC-900, in progress
- Projects — Each project with description, tech stack, link
- Blog/Notes — Key learnings (optional)
- Contact — LinkedIn, GitHub, email
📋 Phase 4 Checklist
- Generative AI: concepts and limitations mastered
- Prompt Engineering: prompt library created
- NIST AI RMF: complete framework studied
- ISO 42001: overview completed
- Responsible AI: impact assessment completed
- Microsoft Copilot: functionality mastered
- Project 1: Security Incident Assistant completed
- Project 2: Risk Assessment Assistant completed
- Portfolio website published
- GitHub portfolio with 7+ projects
🔗 Resume Value
After this phase:
"I developed AI assistants for security incident analysis and risk assessment using Python and Azure OpenAI. I implemented AI Governance frameworks (NIST AI RMF, ISO 42001) and designed Responsible AI processes that ensure fairness, transparency, and accountability in enterprise AI deployments."