Capstone: End-to-End AI Product Workflow

Introduction to the Capstone

This capstone walks you through a complete product development cycle using AI as your co-pilot.

You will pick a real product problem from your work. Then you will work through five steps. Each step produces a deliverable.

You will use skills from all previous modules. By the end, you will have a full product brief ready to share with your team.

Time estimate: 4-6 hours spread over one week.

Step 1: Pick a Product Problem

Choose one feature request, user problem, or product gap you want to solve.

Good problems:

• Something customers have asked for (at least 5 times)
• Something that solves a real pain point you have observed
• Something you can scope in 2-4 weeks of engineering

Avoid:

• Vague problems ("improve user engagement")
• Problems requiring company strategy ("enter new market")
• Problems with unclear success metrics

Step 2: Synthesize Discovery Inputs with AI

Gather raw data about the problem.

Collect:

• Customer support tickets mentioning this problem (5-10)
• Interview notes from users you have talked to (3-5 quotes)
• Feature requests from your CRM or feedback tool (10-15)
• Competitive solutions (how do competitors solve this?)

Now use AI to synthesize.

Prompt: "I am considering a new feature to [problem]. Here is customer feedback:

[Paste 15 pieces of feedback]

Please: (1) Identify 3-4 themes about what users want, (2) Show which feedback supports each theme, (3) Identify any contradictions, (4) Suggest what users care most about based on frequency."

AI will cluster feedback into themes. You now have a clear picture of the problem.

Deliverable: Synthesis document with 3-4 themes, supporting quotes, and analysis.

Step 3: Draft PRD Section with User Stories

Use the themes from step 2 to draft your problem statement and user stories.

Prompt: "Based on the customer feedback I shared, here are the key themes:

[List themes]

Please write: (1) A problem statement (2 paragraphs) that explains what is frustrating users, and (2) 3-4 user stories from different user perspectives."

AI will draft these sections. You review, refine, and add your product strategy.

Deliverable: Problem statement and 3-4 user stories in PRD format.

Step 4: Prioritize Features

Now you need to decide what to build first. Use AI to score options.

If you have multiple related features, list them. Score with RICE or ICE.

Prompt: "I have three potential features to solve this problem:

1. [Feature A]
2. [Feature B]
3. [Feature C]

Context: [Your constraints, team capacity, strategic goals]

Please score each with RICE (Reach, Impact, Confidence, Effort). Which should we build first?"

AI will score them. You validate the scores against your team's estimates.

Deliverable: Prioritization analysis with RICE scores and recommendation.

Step 5: Frame an AI Feature and Design Experiment

If part of your solution involves AI, design the experiment.

If your problem is: "Users do not know what task to work on next," an AI solution could be: "AI suggests next best task based on user patterns."

Design the experiment:

Prompt: "I am considering an AI feature that [solves your problem]. Here is what it would do:

[Description of AI feature]

Please design an experiment to test this. Include: (1) Hypothesis, (2) Method (A/B test, shadow mode, etc.), (3) Success metrics, (4) Timeline, (5) Guardrails."

AI will outline the experiment. You refine it with your team's input.

Deliverable: Experiment brief (1 page) with hypothesis, metrics, and decision rules.

Quality Checklist

Before you finalize, check each deliverable:

Deliverable 1: Synthesis document

• Themes are grounded in actual customer quotes
• Contradictions are flagged
• You validated top 3 themes against raw data

Deliverable 2: Problem statement and user stories

• Problem statement does not assume a solution
• User stories are from real user perspectives you researched
• Each story is independent and small enough to build in one sprint

Deliverable 3: Prioritization analysis

• RICE scores are based on best estimates, not guesses
• You involved engineering in effort estimation
• You explained trade-offs (why this feature over that one)

Deliverable 4: Experiment brief

• Hypothesis is specific and testable
• Success metrics are measurable
• Decision rules are clear (expand / iterate / kill)
• Timeline is realistic

Putting It Together

Your capstone output is a product brief with four sections:

Section 1: Opportunity (synthesis of customer feedback)

Section 2: Solution (problem statement + user stories)

Section 3: Prioritization (why this feature first)

Section 4: Validation Plan (experiment to test the idea)

This brief is something you can share with your team and leadership. It shows you have thought through the problem, listened to customers, and have a plan to validate your idea.

Next Steps

After completing the capstone, bring your brief to your product team.

Use it to:

• Get buy-in from leadership
• Kick off design and engineering
• Set success metrics before launch
• Plan how to measure results

Your capstone brief becomes the foundation for the feature project.

Reflection

After you finish, ask yourself:

1. Did AI help me think more clearly about the problem? How?
2. What surprised me in the customer feedback?
3. What would I do differently next time?
4. What parts of this process can I automate in future projects?

Use these reflections to improve your PM workflow.