Building with ChatGPT and Claude

A Practical One-Day Course for Builders and Technical Teams

ChatGPT and Claude have developed from conversational assistants into practical platforms for software development, technical analysis, document processing, tool integration, and workflow automation. OpenAI provides builder capabilities through the Responses API, tool calling, structured outputs, agent tooling, and Codex. Anthropic provides comparable capabilities through the Claude API, tool use, structured outputs, prompt caching, Claude Code, and Model Context Protocol connectivity.

This one-day course focuses on what technical teams can realistically learn and practise within a single intensive day. Participants will compare both platforms, create effective technical prompts, use AI for development work, make basic API calls, and design a simple workflow that can use either model.

The instructor has more than 30 years of industry experience and will deliver practical, industry-demanded content based on real technical workflows rather than an academic or theory-heavy approach.

Learning Outcomes

Participants will be able to:

Compare ChatGPT and Claude for common technical workloads
Create reusable prompts for analysis, coding, and documentation
Use ChatGPT and Claude to support software-development activities
Make basic OpenAI and Anthropic API requests
Generate structured outputs suitable for applications
Explain tool calling and Model Context Protocol
Design simple routing, fallback, and model-review workflows
Apply basic evaluation, security, and cost controls

Prerequisites

Basic software-development or technical-integration knowledge
Familiarity with APIs and JSON
Basic Python, JavaScript, or TypeScript knowledge
Access to ChatGPT, Claude, and their developer platforms
A laptop with a code editor and development environment

Training Outline

Understanding the ChatGPT and Claude Ecosystems
1. ChatGPT and the OpenAI platform
  1. ChatGPT workspace
  2. OpenAI Responses API
  3. Structured outputs
  4. Tool calling
  5. Codex coding workflows
2. Claude and the Anthropic platform
  1. Claude workspace
  2. Claude Messages API
  3. Structured outputs
  4. Tool use
  5. Claude Code
3. Selecting the Appropriate Platform
  1. Task complexity
  2. Coding requirements
  3. Context requirements
  4. Response quality
  5. Speed and cost
  6. Security requirements
Prompt Engineering for Technical Teams
1. Prompt Structure
  1. Task definition
  2. Technical context
  3. Constraints
  4. Source material
  5. Required output
  6. Quality criteria
2. Context Engineering
  1. Relevant-context selection
  2. Context boundaries
  3. Large-document handling
  4. Conversation history
  5. Sensitive-data exclusion
3. Cross-Platform Prompt Design
  1. Shared prompt templates
  2. Provider-specific adjustments
  3. Output consistency
  4. Prompt versioning
  5. Comparative testing
ChatGPT and Claude for Development Work
1. Requirements and Design
  1. Requirement clarification
  2. User-story refinement
  3. Acceptance criteria
  4. Architecture options
  5. Technical risk identification
2. Coding and Testing
  1. Code generation
  2. Code explanation
  3. Refactoring
  4. Debugging
  5. Unit-test generation
  6. Code review
3. Technical Documentation
  1. README files
  2. API documentation
  3. Runbooks
  4. Troubleshooting guides
  5. Change summaries
4. Coding Agents
  1. Codex workflows
  2. Claude Code workflows
  3. Repository exploration
  4. Controlled file changes
  5. Command execution
  6. Test execution
  7. Human review
Building with the OpenAI and Anthropic APIs
1. API Environment Preparation
  1. API keys
  2. Environment variables
  3. SDK installation
  4. Secret management
2. OpenAI API Fundamentals
  1. Responses API requests
  2. Instructions and inputs
  3. Response extraction
  4. Usage information
  5. Error handling
3. Anthropic API Fundamentals
  1. Messages API requests
  2. System instructions
  3. Content blocks
  4. Response extraction
  5. Error handling
4. Structured Outputs
  1. JSON schemas
  2. Required fields
  3. Output validation
  4. Invalid-response handling
5. Provider Abstraction
  1. Common request format
  2. Common response format
  3. OpenAI adapter
  4. Anthropic adapter
  5. Configuration-based selection
Tools and Connected Workflows
1. Tool-Calling Process
  1. Tool definitions
  2. Input schemas
  3. Model tool selection
  4. Application-side execution
  5. Tool-result submission
  6. Final response generation
2. Model Context Protocol
  1. MCP clients
  2. MCP servers
  3. Tools and resources
  4. External-system connectivity
  5. Authentication and permissions
3. Tool Security
  1. Least-privilege access
  2. Input validation
  3. Tool allowlists
  4. Destructive-action restrictions
  5. Human approval
  6. Audit logging
Combining ChatGPT and Claude
1. Dual-Model Patterns
  1. User-selected provider
  2. Task-based routing
  3. Primary and fallback provider
  4. Generator and reviewer
  5. Parallel response comparison
2. Evaluation
  1. Technical correctness
  2. Instruction adherence
  3. Output consistency
  4. Response latency
  5. Token usage
  6. Cost comparison
3. Production Considerations
  1. Prompt injection
  2. Confidential information
  3. API-key protection
  4. Logging and monitoring
  5. Failure handling
  6. Human verification
  7. Provider lock-in
4. Guided Builder Exercise
  1. Selecting a technical use case
  2. Creating a shared prompt
  3. Calling both providers
  4. Validating structured outputs
  5. Comparing results
  6. Adding provider fallback
  7. Reviewing security and cost

Disclaimer

This outline is provided as a professional guideline for course planning and delivery. The trainer reserves the right to amend, reorganize, expand, reduce, substitute, or omit topics without prior notice where reasonably necessary to accommodate participant experience, available instructional time, technical conditions, platform changes, or organizational requirements, while preserving the course’s principal learning objectives.

Practical, connected learning

My wider training approach brings hands-on implementation and systems thinking together, connecting technology with real operational needs.

Building with ChatGPT and Claude