For product, design and engineering job seekers. Bring a practice brief or a question.
Dave WangProduct Dave
What you'll learn
By the end of this hour
1Plan your approach before opening an AI toolClarify the brief, time limit and permitted AI use, then decide what your response needs to demonstrate.
2Try AI on one part of a practice taskExplore research, a prototype or a draft with AI, then review the output and decide what to keep or change.
3Explain the thinking behind your submissionPractise describing your decisions, trade-offs and AI contribution, and get feedback from a peer.
Lesson map · tap a card to jump · 60 minutes
How the hour runs
From my take-homes A mistake I made
RentSpree: a polished concept that answered a later question
What I built: SteadBuilt around a landlord buying another property.[DAVE: Stead screenshot]🔒 Stead ↗
panel feedback→
What they asked: Owner MomentumHow does an agent hand an owner off after tenant screening, and help that owner complete a first management task?[DAVE: Owner Momentum screenshot]🔒 Owner Momentum ↗
Ask the room: what can a polished take-home establish, and what can't it?
Both RentSpree links need my access.
The line to remember
A prototype is a question you can click.
The quality of that question matters more than the polish of the screens.
From my work Kajabi · how I landed my first US product role
Show, don't tell: 11 product ideas and a Loom
I was building my online courses on Kajabi, and I spotted a PM role there
A friend introduced me to the hiring manager
Then I sent a Notion page with 11 product ideas, and a Loom walking through my thinking
I wrote the ideas as jobs to be done, from my own experience as a creator
Kajabi Creator Tools Product Ideas ▶ Loom
JTBD: Help me create better content"Help me schedule content that needs to be published across multiple platforms"
JTBD: Help me own my content"...so I'm not a slave to the algorithm"
JTBD: Help me own my knowledge products"Help me set up different pricing for different geographic locations"
11 ideas in total
"This wasn't flashy. It was real. It showed I understood their users, business, and space." Nobody set me this task. A cheap artefact that starts a real conversation is a prototype too.
Part 1 · plan before you open an AI tool
Read the task like a reviewer, before you prompt anything
1The briefWhat exactly are they asking you to decide or make? For which user?
2The time limitHow long do they expect you to spend? Plan your hours to it, not past it
3The AI rulesIs AI allowed? For which parts? Do they want you to say how you used it?
4What it must showWhich skills is this task testing? Your response should make those easy to see
If the time limit or AI rules aren't written down, ask before you start.
Part 1 · answer the exact brief
Write the brief in one sentence, then mark the evidence
The one-sentence brief
For [this user], we need to learn whether [this proposed change] helps them complete [this outcome], under [these constraints].
Interview observation · one data point"A childcare director described the coverage problem."
Untested business hypothesis"Directors will pay for this workflow." A prototype can make the workflow concrete. It can't turn this into a fact.
If the brief asks for activation, an elegant expansion concept belongs later. That was the RentSpree lesson.
Keep asking these all session
Five questions to keep asking
1What am I being asked to decide?
2What have I actually observed?
3What must the user complete?
4What would this prototype let me see?
5What would make me change direction?
Try it Part 1 · 5 minutes · no AI yet
Plan your practice brief before you prompt
1Write the brief in one sentenceName the user, job, constraint and finish line
2Write down the time limit and AI rulesNot stated? Write the question you'd send the recruiter
3Mark the evidenceOne thing observed or supplied, and two assumptions
4Decide what your response must showAnd the riskiest assumption it should test
No brief? Use this one
A childcare centre director learns before opening that a teacher is sick and needs safe, approved coverage for the morning.
Part 2 · try AI on one part
Pick one part. Research, a prototype or a draft.
Whichever you pick, you review the output and decide what to keep or change. That part is yours.
From my work Research path · fire off 2 agents
A Business Analyst agent and a Designer agent
1Business Analyst agent to help with market research
Research on segmentation, pain points and customer journey "What other ways to segment the customer base?"
Research on market size and prioritise based on reach and customer value "Help me research on the market size and customer value based on the idea"
Output in high-level point form and in table format
Use the BA agent to create the prompt for the design agent
prompt→
2Designer agent to help with prototyping
Research the brand guidelines and style guide for the company, in preparation for prototyping
Create the prototype based on the prompt
Then ask AI for the quant to verify the segment prioritisation.
From my own PM interview prep notes. [DAVE: which tool you set these agents up in, e.g. a Claude Project or custom GPT]
From my work Draft path · the PRD prompt I use
Make the AI ask you questions before it writes
1Wait modeIt doesn't write anything until I type WRITE THE PRD
2Collect inputs, one by one12 questions: idea, problems, market, segments, JTBD, metrics, features, risks
3Then a fixed structureSame sections every time, so every PRD is easy to compare
PRD writing prompt (excerpt)
Role:
You are an expert Product Manager and PRD writer...
Behavior Rules
1. Wait Mode:
Do not start writing the PRD immediately.
Wait until I explicitly say: "WRITE THE PRD"
2. Collect Inputs First:
Ask me to provide the following information
in sequence (one by one):
1. Product Idea / Concept
2. Business Problem
3. Customer Problem
... 12. Year 1 / Year 3 Goals
Prototype path · the process behind it
Four build steps, with a step before and a loop after
BeforeMy four build stepsAfter
Brief and evidence
→
PRD and assumption
→
One customer outcome
→
Four to five screens
→
Clickable flow
→
Observed task
→
Revise the PRD
↺ and round again with the revised PRD
My four build steps: a good PRD, user goals and steps, the four to five key screens, then stitch them together. My take-homes taught me to add a decision before them and a learning loop after them.
From my take-homes Brightwheel · demonstrated vs simulated
Brightwheel staff coverage
A teacher texts in sick before opening
The director requests cover
A replacement accepts
The director sees the arrival, or an exception
Coverage is verified: the finish line
Simulated: the actions behind each screen. The first release I proposed: narrower, director-approved coordination.
Four take-home prototypes, four things they taught me
1MetaStay close to the charter. Come back to the team's actual problem when a bigger idea pulls you awayCreator Studio ↗
2YoodliMake evidence useful for a decision. A practice score isn't a proven business outcomeInsights ↗
3Tin CanStart with lived use, then narrow. Approvals, calls and the AI coordinator are simulated, and it says soRetention ↗
4RentSpreeAnswer the current job first. A login or an opened message isn't the completion event🔒 Owner Momentum ↗
From my work Prototype path · the prompt from my interview prep
The vibe coding prompt
A prompt I used in my own PM interview prep. The BA agent writes the prompt, then you paste it into the Designer agent.
Vibe coding prompt (BA → Designer)
You are a BA Agent writing a prompt for a Designer agent to help create a clickable prototype for a product interview. Here is the context. Output a prompt I can paste into a design agent.
PRODUCT SOLUTION
- Solution should involve some form of GenAI and [your company's strengths / moats]
USERS & PAIN POINTS
- Copy and paste from Segmentation, JTBD and Pain Points
KEY FLOWS
- List out who the target users are
- The end goal is XYZ (start with the goal)
- Use navigation arrows to link different screens together
- Add assumptions and questions on the side
KEY SCREENS
- Start with 4 to 5 screens first, and give flexibility to expand on the screens
- Each screen should have arrows linking them together from one screen to another
- Add explanation to each of the key features on the screen: customer pain point, product solution, how this ties back to the company goal
- Give the ability to restart the prototype
GUARDRAILS
- Use familiar UI and UX from the brand guideline
- Call out any Error Management and edge cases not in the design
From my workWhat I learnt Prototype path · tools and mistakes
Front end first. The backend can wait.
Lovable→GitHub→Claude Code or Codex
1Front end firstBuilding front and back end together "costs a lot of time, energy, and tokens"
2Penny went offline"I didn't set up a backend for it, which makes sense." We'd only vibe coded the front end
3My rookie mistakeOn Railway, changes don't apply until you press Deploy
When it breaks: take a screenshot, put it into Claude, describe what you expected, and try the flow again.
Prompt you can adapt · works for any of the three paths
Ask for the screen map before the build
Clickable concept prompt
Here is my brief, brain dump, and evidence. Help me create a clickable concept for one target user.
First, restate the exact customer job and the assumption this prototype should test. Separate observed facts from assumptions. Propose a flow of no more than five key screens, including one failure or exception path. For each screen, say what the user sees, decides, and does.
Use fictional data where needed and label simulated behaviour. Do not invent customer quotes, business metrics, integrations, or validation. Show me the screen map before building. Once I choose the scope, build the clickthrough and include a restart.
AI can make your choices visible faster. It shouldn't silently make them for you.
Try it Part 2 · 11 minutes
Try AI on one part, then review it like the hiring team
1Pick one part · 1 minResearch, a prototype or a draft. Check it's allowed by your AI rules
2Prompt · 6 minPaste your one-sentence brief and evidence. Use the prompt on the last slide, or one of mine
3Review · 4 minMark each part: keep, change or cut. Write one line on why
Review checklistDoes it answer the brief I wrote, or a later question? Did it invent quotes, metrics, integrations or validation? Which claims are evidence, and which are assumptions? Is anything simulated, and is it labelled? What would I cut to fit the time limit?
Part 3 · explain the thinking behind your submission
Say the decisions out loud, not just the output
The reviewer can see what you made. They can't see why
Name what you chose, what you left out, and why
Be plain about what AI did and what you decided
Say what you'd test next, and what would change your mind
Two-minute explanation
The brief, in one sentence: ...
What I focused on, and why: ...
What I left out (my trade-offs): ...
What's evidence, and what's an assumption: ...
How I used AI, and what I changed: ...
What I'd test next: ...
From my take-homes Brightwheel · how I explain my AI contribution
AI makes the choices visible faster. The PM makes them.
Where AI helpedAccelerated market and customer research. Organised the director interview. Drafted strategy. Made screens and iterated the clickthrough. Separate strategy and prototype agents worked from a shared project folder.
The calls I madeChose the customer problem. Challenged the generated ideas. Cut the first release. Edited the story.
The PM owns the brief, the customer outcome, assumptions, scope, trade-offs and how feedback is read. The director conversation grounded the work, but it was still one interview.
Peer rubric · can your partner identify all six?
After two minutes, can your partner name all six?
1One target user
2One completed outcome
3The evidence and assumptions
4A coherent path with an exception
5What is simulated
6The decision the test is meant to inform
If a peer can't, revise the story before adding more screens.
Try it Part 3 · peer feedback · 7 minutes
Explain it to a partner. Then swap.
1Explain · 2 minUse the two-minute structure. Your partner stays quiet and listens
2Built a flow?Give your partner a task first, then watch. Observe before explaining
3Feedback · 1 minWhich of the six could they name? Where did they hesitate?
4SwapSame again. Then write one sentence you'd change in your submission
A positive comment about the look of a screen is useful design feedback. It doesn't establish repeat use, willingness to pay, feasibility or a business result.
Debrief · 2 minutes
What changed, and what goes back into your submission?
What did your partner actually say back, or do?
Which assumption became more or less plausible?
Which screen can you remove?
What do you still need to learn from a real customer or engineer?
What sentence goes back into your submission?
Wrap · before you hit submit
Check it the way a reviewer will
The brief, in one sentence, near the top
Inside the time limit, and AI used the way the rules allow
AI use explained, if they asked for it
Simulated or fictional data labelled
Your decisions and trade-offs, in plain words
Every link opened the way a reviewer would
A mistake I made · Wayfair Home ProjectsIt helped me iterate quickly from Saved Boards to a visual project flow. The old public URL now returns 404. Check the final demo link as a reviewer would.
To close
"Tools will keep changing. The product work underneath them is more durable: understand the brief, find the customer job, make one idea tangible, watch someone try it, and change your mind when the evidence tells you to."
"I've built prototypes that were useful because they made a workflow clear. I've also built a polished one that missed the immediate question. Both experiences made the next version better."
Your next take-home
Plan it. Try AI on one part. Explain it.
1PlanOne-sentence brief, time limit, AI rules, and what it must show. Before any prompt
2Try AI on one partResearch, a prototype or a draft. Then keep, change or cut
3ExplainTwo minutes to one person. Can they name the six?
Share what you try:[DAVE: where to share, e.g. Maven community / next session]
