Exercises
30 progressive hands-on exercises covering AI automation, from basic email classification to advanced multi-modal AI systems.
Exercise Capabilities Matrix
| Capability | Symbol | Description | Exercise Count |
|---|---|---|---|
| Classification | CLA | Categorise and label data | 28 exercises |
| Generation | GEN | Create new content | 23 exercises |
| Summarisation | SUM | Extract key insights | 11 exercises |
| Search | SEA | Find and retrieve information | 13 exercises |
| Transformation | TRA | Convert between formats | 10 exercises |
| Vision | VIS | Process visual content | 6 exercises |
| Audio | AUD | Process audio content | 5 exercises |
| Storyboard | STO | Create visual narratives | 7 exercises |
| Video | VID | Generate video content | 5 exercises |
Foundation Track (Exercises 1-13)
Exercise 1: Smart Email Classifier
Duration: 45 min | Difficulty: Beginner | Status: Available Capabilities: Classification Build an autonomous email classification system using AI to categorise and route messages. AI Models: GPT-4o | Key Concept: Agentic Decision Making
Exercise 2: Customer Support Ticket Router
Duration: 60 min | Difficulty: Intermediate | Prerequisites: Ex. 1 Capabilities: Classification, Generation, Summarisation Create intelligent routing for support tickets with auto-responses and prioritisation. AI Models: GPT-4o | Key Concept: Multi-Agent Workflows
Exercise 3: LLM as a Judge - Quality Control System
Duration: 60 min | Difficulty: Intermediate | Prerequisites: Ex. 1 | Status: Available Capabilities: Generation, Classification, Iteration Build an AI quality control system where one LLM generates content and another evaluates it, iterating until quality standards are met. AI Models: Claude 3.5 Sonnet, GPT-4o-mini | Key Concept: Self-Improving AI Workflows
Exercise 4: Connecting Multiple Workflows
Duration: 75 min | Difficulty: Intermediate | Prerequisites: Ex. 3 | Status: Available Capabilities: Workflow Orchestration, Classification, Generation Build an intelligent email response system by connecting three workflows: master orchestrator, email classifier, and LLM-as-a-judge response generator. AI Models: Google Gemini | Key Concept: Workflow Composition & Modular Architecture
Exercise 7: Meeting Notes Summariser
Duration: 45 min | Difficulty: Intermediate | Prerequisites: Ex. 3 Capabilities: Classification, Summarisation, Search Automatically transcribe and summarise meeting recordings with action items. AI Models: GPT-4o, Claude Sonnet | Key Concept: Remote Work Automation
Exercise 9: Customer Feedback Consolidator
Duration: 40 min | Difficulty: Beginner | Prerequisites: Ex. 3, 7 Capabilities: Classification, Summarisation, Search Aggregate and analyse customer feedback from multiple channels. AI Models: GPT-4o, Text-Embedding-3 | Key Concept: Voice of Customer AI
Exercise 13: Language Bridge Translator
Duration: 35 min | Difficulty: Beginner | Prerequisites: Ex. 2 Capabilities: Classification, Generation, Transformation Create real-time multi-language translation workflows. AI Models: GPT-4o, Google Translate API | Key Concept: Global Communication
AI Integration Track (Exercises 4-20)
Exercise 5: Reddit Trend Content Generator
Duration: 50 min | Difficulty: Intermediate | Prerequisites: Ex. 1, 3 Capabilities: Classification, Generation, Search Mine Reddit for trending topics and generate relevant content automatically. AI Models: GPT-4o, Text-Embedding-3 | Key Concept: Social Media Mining
Exercise 5: Personalised Marketing Copy Creator
Duration: 75 min | Difficulty: Advanced | Prerequisites: Ex. 2, 4 Capabilities: Classification, Generation, Search Generate hyper-personalised marketing content based on customer profiles. AI Models: GPT-4o, Claude Sonnet, Text-Embedding-3 | Key Concept: Hyper-Personalisation
Exercise 6: Multi-Platform Content Scheduler
Duration: 60 min | Difficulty: Intermediate | Prerequisites: Ex. 4, 5 Capabilities: Classification, Generation, Transformation Automate content creation and scheduling across multiple social platforms. AI Models: GPT-4o | Key Concept: Cross-Platform Automation
Exercise 8: Research Article Digest
Duration: 50 min | Difficulty: Intermediate | Prerequisites: Ex. 3, 7 Capabilities: Generation, Summarisation, Search Create personalised research digests from multiple sources. AI Models: GPT-4o, Perplexity API | Key Concept: Knowledge Management
Exercise 10: Smart Document Search
Duration: 55 min | Difficulty: Intermediate | Prerequisites: Ex. 3, 8 Capabilities: Classification, Summarisation, Search Implement RAG-based intelligent document search and retrieval. AI Models: GPT-4o, Text-Embedding-3, Pinecone | Key Concept: RAG Implementation
Exercise 12: Customer Query Resolution
Duration: 45 min | Difficulty: Intermediate | Prerequisites: Ex. 10, 11 Capabilities: Classification, Generation, Search Build an agentic support system with knowledge base integration. AI Models: GPT-4o, Text-Embedding-3, Claude Sonnet | Key Concept: Agentic Support
Exercise 15: Content Style Adapter
Duration: 50 min | Difficulty: Intermediate | Prerequisites: Ex. 5, 13 Capabilities: Classification, Generation, Transformation Adapt content to maintain brand consistency across channels. AI Models: GPT-4o, Claude Sonnet | Key Concept: Brand Consistency AI
Exercise 16: Image Description Generator
Duration: 40 min | Difficulty: Beginner | Prerequisites: Ex. 1 Capabilities: Classification, Generation, Vision Generate accessibility descriptions for images using vision AI. AI Models: GPT-4o-Vision, BLIP-2 | Key Concept: Accessibility Automation
Exercise 17: Visual Content Moderator
Duration: 55 min | Difficulty: Intermediate | Prerequisites: Ex. 16 Capabilities: Classification, Generation, Vision Automatically moderate visual content for safety and compliance. AI Models: GPT-4o-Vision, CLIP, Moderation API | Key Concept: Content Safety AI
Exercise 19: Voice Note Transcriber & Processor
Duration: 45 min | Difficulty: Intermediate | Prerequisites: Ex. 7 Capabilities: Classification, Summarisation, Audio Convert voice notes to actionable tasks and summaries. AI Models: Whisper-v3, GPT-4o | Key Concept: Voice-First Workflows
Exercise 20: Podcast Content Extractor
Duration: 60 min | Difficulty: Intermediate | Prerequisites: Ex. 19, 8 Capabilities: Classification, Generation, Summarisation, Audio Extract key insights and generate social content from podcasts. AI Models: Whisper-v3, GPT-4o, Claude Sonnet | Key Concept: Audio Content Mining
Advanced Workflows Track (Exercises 11-23)
Exercise 11: Code Documentation Explorer
Duration: 65 min | Difficulty: Advanced | Prerequisites: Ex. 8, 10 Capabilities: Classification, Generation, Search Generate and search code documentation using AI. AI Models: GPT-4o, CodeT5, Text-Embedding-3 | Key Concept: Developer Productivity
Exercise 14: Data Format Harmoniser
Duration: 70 min | Difficulty: Advanced | Prerequisites: Ex. 13 Capabilities: Classification, Generation, Transformation Transform and harmonise data across different enterprise systems. AI Models: GPT-4o, Claude Sonnet | Key Concept: Enterprise Integration
Exercise 18: Product Catalogue Analyser
Duration: 75 min | Difficulty: Advanced | Prerequisites: Ex. 16, 17, 10 Capabilities: Classification, Generation, Search, Vision Analyse and categorise products using vision and text AI. AI Models: GPT-4o-Vision, CLIP, Text-Embedding-3 | Key Concept: E-commerce Automation
Exercise 21: Multi-language Audio Support
Duration: 50 min | Difficulty: Advanced | Prerequisites: Ex. 19, 13 Capabilities: Classification, Generation, Transformation, Audio Build global voice support with real-time translation. AI Models: Whisper-v3, GPT-4o, Google Translate | Key Concept: Global Voice Support
Exercise 22: Visual Storyboard Creator
Duration: 75 min | Difficulty: Advanced | Prerequisites: Ex. 16, 8 Capabilities: Classification, Generation, Search, Vision, Storyboard Generate visual storyboards from text descriptions. AI Models: GPT-4o, DALL-E 3, CLIP | Key Concept: Creative AI Workflows
Exercise 23: Marketing Storyboard Automation
Duration: 80 min | Difficulty: Advanced | Prerequisites: Ex. 22, 5 Capabilities: Classification, Generation, Search, Storyboard Automate campaign visualisation with AI-generated storyboards. AI Models: GPT-4o, DALL-E 3, Text-Embedding-3 | Key Concept: Campaign Visualisation
Industry Applications Track (Exercises 24-30)
Exercise 24: Educational Content Storyboard
Duration: 70 min | Difficulty: Advanced | Prerequisites: Ex. 22, 15 Capabilities: Classification, Generation, Transformation, Storyboard Create educational visual content with AI-powered storyboards. AI Models: GPT-4o, DALL-E 3, Claude Sonnet | Key Concept: EdTech Automation
Exercise 25: RunwayML Product Demo Videos
Duration: 90 min | Difficulty: Expert | Prerequisites: Ex. 22, 18 Capabilities: Classification, Generation, Search, Vision, Storyboard, Video Generate product demo videos using AI video generation. AI Models: GPT-4o, DALL-E 3, RunwayML Gen-3 | Key Concept: AI Video Generation
Exercise 26: Pika Labs Training Videos
Duration: 85 min | Difficulty: Expert | Prerequisites: Ex. 21, 24 Capabilities: Classification, Generation, Transformation, Audio, Storyboard, Video Create corporate training videos with AI voice and video. AI Models: GPT-4o, ElevenLabs, Pika Labs 1.5 | Key Concept: Corporate Video AI
Exercise 27: Social Media Video Pipeline
Duration: 80 min | Difficulty: Expert | Prerequisites: Ex. 25, 26 Capabilities: Classification, Generation, Transformation, Storyboard, Video Build a viral content factory with AI video generation. AI Models: GPT-4o, RunwayML Gen-3, Pika Labs 1.5 | Key Concept: Viral Content Factory
Exercise 28: Reddit Trend-to-Video Pipeline
Duration: 120 min | Difficulty: Expert | Prerequisites: Ex. 4, 27, 10 Capabilities: Classification, Generation, Summarisation, Search, Transformation, Vision, Storyboard, Video Transform trending Reddit content into engaging videos. AI Models: GPT-4o, DALL-E 3, RunwayML Gen-3, CLIP, Text-Embedding-3 | Key Concept: Trending Content Automation
Exercise 29: AI Agent Customer Journey
Duration: 90 min | Difficulty: Expert | Prerequisites: Ex. 2, 12, 10 Capabilities: Classification, Generation, Summarisation, Search, Transformation Build autonomous customer experience with multi-agent systems. AI Models: GPT-4o, Claude Sonnet, Text-Embedding-3, AutoGPT | Key Concept: Autonomous Customer Experience
Exercise 30: Omnichannel AI Support Bot
Duration: 120 min | Difficulty: Expert | Prerequisites: Ex. 12, 21, 18 Capabilities: Classification, Generation, Summarisation, Search, Transformation, Vision, Audio Create multi-modal support across all communication channels. AI Models: GPT-4o, Whisper-v3, GPT-4o-Vision, Claude Sonnet | Key Concept: Multi-Modal Support
Learning Paths
Beginners: Start with exercises 1, 3, 9, 13, 16 to build foundational skills.
Intermediate: Progress through exercises 2, 4, 6, 7, 8, 10, 12, 15, 17, 19, 20.
Advanced: Work through exercises 5, 11, 14, 18, 21, 22, 23, 24.
Expert: Complete exercises 25, 26, 27, 28, 29, 30 for multi-modal AI systems.
Recommended Sequences
Exercise 1 is the starting point. Choose a path based on your interests:
- Support & Service: 1 → 2 → 12 → 29 → 30
- Content Creation: 1 → 3 → 4 → 5 → 6 → 27
- Workflow Orchestration: 1 → 3 → 4 → (advanced multi-workflow systems)
- Visual AI: 1 → 16 → 17 → 18 → 22 → 25
- Audio Processing: 3 → 7 → 19 → 20 → 21 → 26
- Enterprise Integration: 2 → 13 → 14 → 15 → 24