You've documented what you discovered about your partner. But how much of what you "know" is actually assumption?
Before you start building anything, surface the assumptions you're carrying — because building on wrong assumptions wastes everyone's time.
How This Works
The AI challenges each assumption you list. It won't let you off easy:
The AI will push you with moves like:
"That's not an assumption — that's a fact. What are you ASSUMING about it?"
"What evidence do you actually have for this vs. what just feels true?"
"What changes about your solution if this assumption is wrong?"
"You listed 3 assumptions. Most people stop too early. What are you assuming about your partner that you haven't examined?"
What You'll Produce
For each of 5+ assumptions about your partner's problem:
What's the assumption?
What evidence do I have? (specific — from the conversation, not guessing)
What could be wrong about this?
What changes about my approach if this assumption is false?
📋 What Good Looks Like
Before you start — here's what a strong assumption audit entry looks like
Fictional example based on the study problem partner. Yours should reflect YOUR partner's situation.
Assumption: My partner needs a better study planner. Evidence I have: She said she "wastes time figuring out what to study." She described going back and forth between Canvas tabs. What could be wrong: Maybe she already has a planner and the problem is she doesn't trust it. Maybe the real issue is decision anxiety — she's afraid of picking the wrong thing to study. A planner wouldn't fix that; it might make it worse by showing her MORE options. What changes if I'm wrong: If it's anxiety, not organization, then building a planner actively makes things worse (more decisions). I'd need to build something that REMOVES choices, not organizes them.
Assumption: She would use a digital tool (app, website). Evidence I have: She had her laptop open during our conversation. She mentioned Canvas multiple times. What could be wrong: She also said "I sometimes write things down" — maybe she's a paper-first person and a digital tool adds friction. I assumed digital because I'm a CS major. That's MY bias, not something from the conversation. What changes if I'm wrong: I'd need to design for paper or a hybrid approach. Ask her directly before committing to building an app.
Notice how the second assumption is about the SOLVER's bias, not just the problem. The best audits include assumptions about yourself, not just your partner.
AI-Guided Discussion
Use your preferred AI chatbot (ChatGPT, Claude, etc.) to surface and stress-test your assumptions. The AI won't accept surface-level answers — it will push you to distinguish between evidence and intuition.
Paste the following prompt to start your session:
I'm working on a project for my partner in CST395. I need to surface at least 5 assumptions I'm carrying about their problem before I start building anything. Help me by asking about my partner's situation, then challenge each assumption I list. For each one, push me on: (1) what specific evidence I have from our conversation, (2) what I'm just guessing, and (3) what changes about my approach if this assumption is wrong. Don't let me stop at easy assumptions — push me toward the ones I haven't examined yet.
Start by asking me: "Describe the problem you're solving for your partner and what you learned in your discovery conversation."
What to Submit
Complete the interactive activity and download your responses as JSON.
Format: JSON file download from AI-discussion activity Due: Points: 5