Exploring GenAI with VAILL: Day 8 💬
A 10-day journey into human-centered GenAI for legal professionals
💬 Day 8: Advanced Prompting Techniques
Welcome back! Now that you've developed a strong foundation in GenAI usage and ethical considerations, let's elevate your skills with advanced prompting techniques. Today, we'll explore sophisticated ways to communicate with GenAI tools that can help you work more effectively with the technology you use every day.
Today's Focus: Practice more advanced prompt engineering techniques to get more nuanced, practical, and useful responses from conversational GenAI tools while maintaining professional standards.
🎯 Learning Objectives
Explore chain-of-thought prompting for detailed technology analysis
Learn to use role-based prompts for different perspectives
Develop skills in context manipulation for better results
🔑 Key Concepts
Advanced prompting is about guiding GenAI tools to provide more thorough and useful responses. Think of it as the difference between asking a tech-savvy colleague "How do I use this software?" versus having a structured conversation about your specific needs and challenges.
At VAILL, we've identified three key advanced techniques for upping your prompting game:
Chain-of-Thought Prompting:
Breaking analysis into clear steps
Asking AI to explain its reasoning
Building understanding progressively
Role-Based Prompting:
Looking at issues from multiple perspectives
Understanding different user needs
Discovering varied approaches
Context Manipulation:
Adjusting scenarios to reveal new insights
Exploring different situations
Finding solutions for specific circumstances
Remember: These advanced techniques are tools to help you get better results from AI. They work best when applied to real-world situations you encounter in your daily work.
💡 Pro Tips
Document useful prompts for future reference
Build on successful approaches
Share discoveries with colleagues
Keep ethical guidelines in mind
✍️ Today's Exercise
Let's practice advanced prompting techniques to explore how we can better use the everyday technology tools in our legal work—from email to document management systems to research databases.
Chain-of-Thought Practice: Start by thinking about a technology tool you use daily in your legal work—such as your email system, document management system, research database, or a new-to-you tool that you have yet to fully explore. Then, try this prompt with your chosen tool:
"I want to improve how I use [your chosen tool] in my work. Please act as a workflow consultant and help me: First, identify the main ways I probably use this tool; then suggest less obvious features that could be helpful; show me how these features could save time or reduce errors; and help me think through potential challenges or pitfalls. Please explain your reasoning at each step so I understand why you're making these suggestions."
Role-Based Analysis: Next, let's examine your tool from different perspectives. Ask the AI to analyze your tool:
From an efficiency perspective: What features help work get done faster and reduce repetitive tasks?
From an accuracy perspective: What features help catch mistakes and better organize information?
From a collaboration perspective: What features help share information and improve communication?
[Insert your customized perspective]
Ask the AI to explain why each perspective matters and how it might change how you use the tool.
Context Manipulation: Explore how to use your chosen tool in different situations by asking the AI to address the following contexts:
During a busy day with many urgent tasks
When working on a detailed, complex project
When training someone else on the system
[Insert a context specific to you and your goals in using the tool]
In all of these situations, you can also ask ChatGPT “What else do you need to know, to respond to my query?” Remember that this is a conversation and you can keep asking questions and providing more context specific to your needs and goals to achieve greater insight.
This exercise helps you:
Discover new ways to use familiar tools
Create more efficient workflows
Reduce common frustrations
Share knowledge with colleagues
Remember, the goal isn't just to learn about technology—it's to practice these advanced prompting techniques in ways that immediately improve your daily work. Feel free to try this approach with any technology tool you use in your legal work.
🔍 Common Mistakes to Avoid
Forgetting that this is a conversation and stopping with the first (or even second) output the AI offers
Making prompts too complicated or technical (but there are exceptions to this—see Additional Resources 👇 )
Overlooking opportunities to share discoveries with colleagues
⚡️ Lightning Round (Optional)
Take another technology tool you use regularly and:
Create three different prompts using each technique we learned today
Compare which technique gave you the most useful insights
Share your best discovery with a colleague
Keep notes on which approaches work best for different types of questions
🔄 Quick Recap
Advanced techniques help get better results from AI
Chain-of-thought prompting reveals helpful details
Role-based prompts offer new perspectives
Context manipulation helps solve specific problems
📔 Reflect
Before moving on, spend a few minutes in your learning journal noting:
Which of today's advanced prompting techniques feels most relevant to your highest-value work? Why?
How might you adapt these techniques to your specific needs?
What support or resources would help you implement these ideas?
These final reflections will be particularly valuable as you develop your personal AI implementation plan.
📚 Additional Resources
Database: AI Prompt Library for Law: Research, Drafting, Teaching, and More (AI Law Librarians blog) - “a place for law librarians (and the legal community at large) to share and collect useful prompts,” created by the AI Law Librarians bloggers
Prompting by a master: Greg Siskind’s prompting post on LinkedIn - Greg is an immigration lawyer and creator of the GenAI platform Gen for immigration lawyers; in this post, he shares a very detailed prompt he used to get instant answers in Gen to the question of whether a country and a particular occupation are on the 2024 J-1 Skills List. Read Greg’s post and the prompt (4.5 pages!) to get a glimpse into high-level prompting to achieve a specific, sophisticated purpose using a GenAI tool combining an LLM with relevant legal data.
Interactive: Generative AI for Legal Professionals (Coursera) - Take a deeper dive into GenAI and prompting with VAILL’s founding co-director Mark Williams
👉 Coming Tomorrow
Get ready to compare different AI tools! Next, we'll explore how to choose the right tool for specific tasks and develop strategies for using multiple tools effectively.
Keep experimenting with these advanced techniques, and share your discoveries with us at vaill@vanderbilt.edu!