AI Prompt Engineering
Google AIPrompt Engineering
Master the T.C.R.E.I. framework and transform your basic prompts into powerful AI instructions
The T.C.R.E.I. Framework
A systematic approach to creating effective AI prompts
Task
Define what you want AI to do
Context
Provide background information
References
Give specific examples
Evaluate
Assess AI output quality
Iterate
Refine and improve prompts
Memory Aid
🦀 Tiny Crabs Ride Enormous Iguanas 🦎
Remember the framework with this fun mnemonic!
Iteration Methods
Strategies for improving your prompt outputs
🔄 Revisit Framework
Go back to T.C.R.E.I. steps and improve each element
✂️ Shorter Sentences
Break complex requests into focused, manageable parts
🔄 Different Phrasing
Try alternative wording or analogous tasks
🚧 Add Constraints
Introduce specific limitations to guide AI focus
AI Agents & Multimodal Capabilities
Specialized AI assistants and advanced interaction methods
AI Agents are specialized experts designed to help with specific tasks and answer questions within defined domains.
AI models like Gemini can process multiple types of data simultaneously for richer, more contextual interactions.
Prompt Examples
See how basic prompts transform using the T.C.R.E.I. framework
Basic: suggest a gift related to anime for my friend's birthday.
+ Persona: Act as an anime expert to suggest an anime gift for my friend's birthday.
+ Context: Act as an anime expert to suggest a gift related to anime for my friend's birthday and she is turning 29 years old. Her favourite anime is Shangri-la Frontier.
Task: Create a table that tracks weekly staffing.
Context: I have 10 employees. Create columns for day, name, and shift.
Constraints:
- 3 employees for morning, 2 for afternoon shift
- No morning shift after afternoon shift
- Equal total shifts per week per employee
Persona: Act as a career development training simulator.
Task: Help interns master interview skills.
Stop Rule: Continue until "JAZZ HANDS."
Challenges & Responsible AI Use
Understanding AI limitations and ethical deployment principles
🎭 Hallucination
AI provides inconsistent, incorrect, or nonsensical outputs
⚖️ Biases
AI inherits biases from human-generated training data
🔗 Prompt Chaining
Interconnected prompts for complex tasks
🧠 Chain of Thought
Step-by-step reasoning process
🌳 Tree of Thought
Multiple reasoning paths exploration
🎯 Meta Prompting
High-level behavior instructions