GENERATIVE AI WITH CLAUDE

Get a clear, practical foundation in Claude AI and how it fits into professional workflows, before you write a single line of code.

• What Claude AI is and how it fits into the Anthropic ecosystem
• Claude Haiku, Sonnet, and Opus: Understanding the model family and choosing the right model for your task
• How Claude processes information: Context windows, tokens, and what that means for your workflows
• Temperature settings: When to use precise outputs vs creative drafting
• Writing effective system prompts that define Claude's professional role, output format, tone, and constraints
• Lab: Write and test a system prompt prototype for your own job role in Claude.ai

Move beyond the chat interface. Learn how to access Claude programmatically and build the foundation for automated, integrated workflows.

• Setting up API authentication and making your first successful API call
• Core API parameters: Model, max_tokens, temperature, system, and messages
• Structuring API requests to produce reliable, consistent, professional-grade outputs
• Technical specification writing: How to document what your API tool does, its inputs, outputs, and dependencies
• Lab: Build a working API call for a daily work task and document it as a formal technical specification

Tool use is what transforms Claude from a conversational assistant into an agent. Learn how to define tools, pass them to Claude, and build the interaction model around them.

• What tool use is and why it changes what Claude can do in a work context
• Defining tools as JSON schemas and passing them to the Anthropic API
• How Claude decides when and how to call a tool, and how to handle multi-turn tool conversations
• Interface design for agent interactions: Input schemas, output formats, feedback loops, and error handling
• Lab: Give Claude two tools and build a functional prototype for a real work task

Shift from building single API calls to designing full agent systems. Map out how Claude can pursue a goal across multiple steps — and plan your own workflow to build on Day 2.

• The difference between a chatbot and an agent in a professional work context
• The agent loop: Goal, plan, act, check, repeat
• Identifying real work tasks from your job that are suitable for an agent workflow
• The workflow blueprint template: Trigger, goal, steps, tools, data inputs and outputs, decision points, human-in-the-loop checkpoints, and success criteria
• Lab: Draft a complete workflow blueprint for the Claude agent you will build on Day 2

Apply interaction design principles to make your agent workflow usable, trustworthy, and effective, then build it.

• Core interaction design principles for AI agents: visibility, feedback, user control, error recovery, consistency, and affordance
• Usability testing concepts for AI tools: task completion rate, error rate, time on task, satisfaction, and cognitive load
• Running structured stakeholder review sessions and acting on the feedback you receive
• Building a multi-step workflow by chaining Claude API calls, managing state across steps, and implementing error handling
• Sequential and parallel workflow patterns, and how to add human-in-the-loop checkpoints before irreversible actions
• Lab: Build a working 3-step Claude workflow based on your approved blueprint, tested end-to-end with real data