Episode 1: From Skeptic to Champion: How to Explain AI to an Experienced Evaluator
From Skeptic to Champion
How to explain AI to an experienced evaluator without sounding like someone who just discovered their first TED Talk.
⚔️ A Challenge Awaits
Dr. Patricia Okonkwo has crossed her arms and is giving you The Look™. One wrong word and she's gone back to her pivot tables. Can you win her over?
You're in a staff meeting when AI comes up. Patricia, who has survived every tech fad since the fax machine, slowly leans back. She crosses her arms. She exhales through her nose in a way that could power a small turbine.
"Another buzzword. Tell me — what does it actually DO? And why in the world should I trust it with my data?"
The room goes quiet. Someone's coffee cup makes a very loud noise. All eyes turn to you.
Patricia is waiting. Her coffee is getting cold. Choose wisely.
The Research Assistant Analogy — Unpacked
Patricia's arms slowly uncross. Her eyebrows lift. She puts down her coffee. "So… like a power tool for my existing skills?"
Yes, Patricia. Exactly like that. And now she's going to be the one reminding everyone else about ethical safeguards.
How Priya actually prepared for this conversation
via ChatGPTBefore the meeting, Priya gave ChatGPT a quick brief on Patricia's background — then asked it to meet her there.
"Create a simple analogy comparing AI to a research assistant for an evaluator comfortable with Excel, SPSS, and qualitative coding."ChatGPT generated the research assistant analogy. Yes, AI wrote its own sales pitch. The irony was not lost on Priya.
Generic = forgettable. Priya connected the analogy to tasks Patricia does every week.
"Now connect this to coding interviews, analyzing surveys, and drafting evaluation reports."Priya offered to take one of Patricia's recent interview transcripts and code it live in front of her. Nothing converts a skeptic like watching their own work transform in real time.
Priya didn't wait for Patricia to raise concerns — she got ahead of them.
"What ethical considerations should I mention when using this analogy with a senior evaluator?"- AI lacks contextual and cultural understanding
- Human oversight isn't optional — it's the whole point
- Participants should know when AI is involved in analysis
- Unchecked, AI can amplify existing biases (so check it)
Built on Existing Knowledge
Evaluators know what research assistants do. The analogy doesn't ask Patricia to learn something new — it asks her to recognise something familiar.
Admitted the Limits Upfront
Including what AI can't do — context, culture, judgment — built more trust than any capability list ever could.
Ethics Were Structural
The analogy contains safeguards by design. AI is an assistant, not an authority. That framing sticks.
Kept Her in Charge
"You're still the expert" is the most important sentence. Nobody buys a tool that threatens to replace them.
🚀 Your Turn — Try the Prompt Yourself
Works in ChatGPT, Claude, Gemini, or any LLM that can read a full sentence
Create a simple analogy to explain AI to an experienced evaluator who is skeptical. They're comfortable with Excel, SPSS, and qualitative coding. The analogy should: 1. Connect to tasks they already do 2. Acknowledge AI's limitations honestly 3. Include ethical considerations they'll respect 4. Be memorable and easy to pass on to others Then, based on this analogy, create: - A 30-second elevator pitch for a rushed hallway conversation - A one-paragraph explanation for a program manager who controls the budget - Answers to 3 likely questions a 20-year veteran skeptic will definitely ask
💡 Pro tip: Save the analogy that resonates most with you. You'll use it again and again — and so will the Patricia you convert.
Ready to Go Further?
Both courses are built for practicing evaluators who want to use AI responsibly — not just talk about it at conferences.
AI for International Development & Humanitarian Practitioners
Specialised for development and humanitarian contexts
Explore Course →Already enrolled? Continue your learning journey →
Your job is to give Priya's response. Choose carefully — Patricia has a finely tuned nonsense detector built from two decades of evaluating programs that promised to "change everything."