Capstone Project: Design an AI System

This is your final project. You'll design a comprehensive AI system for a real business — combining strategy, tool selection, workflow design, agent creation, and ROI analysis.

The Assignment

Choose one of the following business scenarios (or use your own business):

Scenario A: The Growing E-Commerce Brand

A direct-to-consumer brand selling handmade skincare products. 15 employees, $2M annual revenue, growing 30% year-over-year. Pain points: drowning in customer emails, can't keep up with content creation, inventory forecasting is manual and often wrong.

Scenario B: The Professional Services Firm

A 25-person consulting firm specializing in environmental compliance. Pain points: proposal writing takes 20+ hours per proposal, knowledge is trapped in individual consultants' heads, client reporting is manual and inconsistent.

Scenario C: The Local Restaurant Group

A group of 4 restaurants with 60 total employees. Pain points: scheduling is a nightmare, customer feedback is ignored, menu optimization is based on gut feeling, marketing is inconsistent across locations.

Scenario D: Your Own Business

Use your real business. This produces the most valuable output — you'll actually use what you create.

What to Deliver

Your capstone project should include all of the following components:

1. AI Readiness Assessment (10%)

• •Current state audit: What tools and processes exist today?
• •Team capability assessment: What's the AI skill level?
• •Data inventory: What data is available and where does it live?
• •Gap analysis: What's missing?

2. Opportunity Map (15%)

• •List 10+ AI use cases for this business
• •Score each on Impact (1-5) and Feasibility (1-5)
• •Plot on the prioritization matrix
• •Select the top 3 for your roadmap

3. Phased Roadmap (15%)

• •Phase 1: Quick wins (Month 1-2)
• •Phase 2: Team workflows (Month 3-4)
• •Phase 3: Integration (Month 5-6)
• •Include specific milestones and success criteria for each phase

4. Detailed Design for Phase 1 (30%)

For your #1 priority initiative:

• •Agent design: System prompt (using CRISP framework), knowledge base outline, tool list
• •Workflow design: Trigger, steps, data flow, error handling, monitoring
• •Tool selection: Evaluation scorecard with 2-3 options compared
• •Implementation plan: Step-by-step build guide with timeline

5. ROI Analysis (15%)

• •Value created (time savings, error reduction, revenue increase)
• •Total cost (tools, setup, maintenance, training)
• •ROI calculation and payback period
• •Sensitivity analysis: What if costs are 50% higher? What if savings are 50% lower?

6. Risk Assessment & Mitigation (15%)

• •Technical risks (AI accuracy, integration issues)
• •People risks (adoption, resistance, skill gaps)
• •Business risks (cost overrun, vendor dependency)
• •Mitigation plan for each risk
• •Kill criteria: When to stop

Quality Standards

Your project will be evaluated on:

• •Specificity: Real numbers, real tools, real timelines (not "some AI tool" or "significant savings")
• •Feasibility: Could a non-technical business owner actually implement this?
• •Completeness: All 6 components addressed thoroughly
• •Critical thinking: Risks acknowledged, limitations noted, honest ROI projections
• •Communication: Clear enough for a business stakeholder to understand and act on

How to Approach This

1. 1.Start with the business problem — not the technology. What's broken? What hurts?
2. 2.Talk to people (or imagine talking to them) — What do employees actually do all day? Where do they waste time?
3. 3.Be realistic — A $2M business isn't going to spend $50K/year on AI tools. Scale your recommendations.
4. 4.Think end-to-end — Don't just design the AI part. Design the human workflow around it.
5. 5.Include the failures — Show that you've thought about what could go wrong.