Three one-hour sessions for UFT staff — designed for all skill levels, packed with real materials you can use right away.
The foundation session. Understand how AI language models actually work, learn the four building blocks of any great prompt, and explore the differences between ChatGPT, Claude, and Gemini so you can choose the right tool for the right job.
Full presentation — 10 slides covering prompt anatomy and LLM differences
Printable reference card with the 4-part prompt framework and examples
Side-by-side feature comparison of ChatGPT, Claude, and Gemini
Run-of-show, timing, scripts, and discussion prompts for session facilitators
How large language models process and respond to prompts
The four elements of every effective prompt: Role, Context, Task, Format
Key differences between ChatGPT (GPT-4), Claude, and Gemini
How to choose the right AI tool based on your task type
Your secret weapon for consistent, on-brand AI results. Learn to build a reusable persona that tells the AI exactly who it is, what it knows, and how to communicate — for your specific UFT role. Includes a complete worked example for UFT Membership and instructions for setting up personas in both GPT-4 and Claude.
10-slide presentation — what personas are, why they work, and the UFT Membership example
Step-by-step persona setup in GPT-4 and Claude, plus fill-in-the-blank template and comparison table
Ready-to-copy persona for UFT Membership Coordinators — paste into any AI session
Fill-in-the-blank persona template for any UFT department or role
What an AI persona is and how it differs from a regular prompt
The 5 elements of every effective persona: Role, Knowledge, Audience, Tone, Goals/Guardrails
How to save a persona in ChatGPT (Custom Instructions & Custom GPTs) and Claude (Projects)
Build a persona specific to your own UFT department and role
The capstone session — put everything together in a gamified challenge. Participants pick a real task from their own job and write the same prompt four times, adding a new expert ingredient each round. Ends with a community vote on the best AI response. Points, levels, and a winner.
Level 1: The Bare Ask (10 pts) → Level 2: + Context (25 pts) → Level 3: + Role & Format (40 pts) → Level 4: + Persona & Constraint (50 pts + vote bonus). Max score: 125 points.
8-slide presentation — rules, all 4 level challenge cards, and debrief discussion guide
Printable challenge card — all 4 levels with write-in boxes and scoring tracker
Print-ready PDF version of the challenge card — US Letter, 2 pages
All 4 levels built out for a UFT CTE role — color-coded to show exactly what each upgrade adds
Print-ready PDF of Sade's full 4-level prompt progression — 8 pages
Minute-by-minute run of show, scripts, troubleshooting, debrief questions, and scoring
Display-ready scoreboard to track names, levels, bonuses, votes, and totals in real time
How adding context transforms a generic AI response into a relevant one
Why format instructions make AI output immediately usable without editing
How a persona makes the AI sound like YOUR department, not a generic bot
How to build a complete expert-level prompt from scratch for any real work task