PM Dashboard with products and goal
PM Dashboard with Backlog grooming
PM Dashboard with Manage Resources
Login page

PM Assistant Dashboard

Inspiration

Product managers waste 40% of their time jumping between Jira, Workday, and Slack just to gather basic information. We watched a PM say, "I spend more time being a data gatherer than an actual product manager."

We asked: What if one AI-powered dashboard could replace hours of manual work?

What it does

A unified platform that brings together everything PMs need:

Products & Goals

Real-time product tracking with progress bars
Migration planning with impact assessments
Query performance metrics showing current load vs. system capacity

Manage Resources

Team overview with availability, skills, and velocity from Workday
AI Resource Matching: Type "build a React dashboard" → instantly finds available developers with React skills

Backlog Grooming

Consolidated backlog across all products with clarity scores
AI Analysis: Click "AI Assist" on any ticket to get:
- Clarity gap identification
- Missing acceptance criteria
- Blocker analysis and unblock strategies
- Action items with specific owners
- Timeline estimates with confidence levels

How we built it

Tech Stack:

Frontend: Next.js 14 + React 18 + Tailwind CSS
Backend: Next.js API Routes (Node.js)
AI: OpenRouter API (Model: meta llama3)
Deployment: Vercel Edge Network

Build Process:

Designed three core tabs (Products, Resources, Backlog)
Integrated AI for resource matching and backlog analysis
Built professional login screen simulating Jira connection
Optimized AI prompts achieving 73% token reduction
Deployed with CI/CD pipeline to Vercel

Challenges we ran into

1. Textarea Focus Bug

Problem: Lost focus after typing one character
Solution: Fixed React re-rendering by restructuring component architecture

2. API Cost Explosion

Problem: Claude Sonnet costs (3 per 1M tokens
Solution: Switched to lamma3 model at )0.10 per 1M tokens

Cost per analysis:

$$ \text{Cost} = \frac{\text{tokens used}}{1,000,000} \times \text{price per 1M tokens} $$

Result: From (0.0045 → )0.00008 per analysis = 98% cost reduction

3. Environment Variables

Problem: API keys failed in Vercel production
Solution: Configured in Vercel dashboard

4. AI Response Formatting

Problem: Markdown blocks breaking JSON parsing
Solution: Sanitize before parsing: javascript let cleanData = aiResponse.replace(/json\n?/g, '').trim(); ```

Accomplishments that we're proud of

🤖 AI Backlog Analysis - Transforms vague tickets into action plans in 10 seconds
💰 98% Cost Optimization - Smart model selection
⚡ 10 hours saved per PM/week - ($500K+ annual gains
🎨 Production-ready UI - Professional design
🚀 5 days to deploy - Concept to production