8-6a: AI Project Dashboard Guide

A great dashboard gives you instant visibility into your AI-driven development pipeline. Here's how to configure the key widgets used in the AI-First Software Project.

Project Dashboard: Q1 Roadmap

Sprint Velocity

42 pts

+15% vs last sprint

Test Coverage

89%

425 tests passed

Deploy Frequency

12/wk

Avg 2.3d lead time

Avg Cost/Feature

$1.12

ROI 357x

Pipeline Status

Planning4

Development6

Testing & Review5

Deployed15

AI vs Human Contribution

75%

AI Generated

25%

Human Reviewed

Based on line-of-code attribution

• 1. Pipeline Status (Bar Chart)
• 2. Sprint Velocity (Number with Trend)
• 3. AI vs Human Contribution (Pie Chart)
• 4. Test Health (Gauge)
• 5. Cost Per Feature (Number)
• 6. Iteration Efficiency (Line Chart)
• Best Practices for Dashboards

---

1. Pipeline Status (Bar Chart)

Purpose: See where every feature sits in the lifecycle.

Configuration:

• Widget Type: Bar Chart (Horizontal)
• Data Source: Main Project Taible
• Group By: Status column (Planning, Dev, Review, Deployed)
• Metric: Count of Rows

Why it matters: Instantly spot bottlenecks. If 15 tasks are in "Review", you need more human reviewers.

---

2. Sprint Velocity (Number with Trend)

Purpose: Track team output speed.

Configuration:

• Widget Type: KPI / Big Number
• Data Source: Main Project Taible
• Filter: Status is "Deployed" AND Deployment Date is "This Sprint"
• Metric: Sum of Story Points
• Trend: Compare to "Previous Sprint"

Why it matters: Measures the raw productivity impact of AI augmentation.

---

3. AI vs Human Contribution (Pie Chart)

Purpose: Understand how much work AI is doing vs humans.

Configuration:

• Widget Type: Pie / Donut Chart
• Data Source: Main Project Taible
• Metrics:
  1. Sum of AI Token Usage (proxy for AI work)
  2. Sum of Human Hours (manual work)
• Note: You may need a computed column to normalize these values into a "Contribution Score".

Why it matters: Proves the ROI of the system to stakeholders.

---

4. Test Health (Gauge)

Purpose: Ensure speed isn't killing quality.

Configuration:

• Widget Type: Gauge
• Data Source: Main Project Taible
• Metric: Average of Test Coverage % column
• Ranges:
  • 0-50% (Red)
  • 50-80% (Yellow)
  • 80-100% (Green)

Why it matters: AI generates a lot of code fast. You must ensure it's tested code.

---

5. Cost Per Feature (Number)

Purpose: Track AI API costs.

Configuration:

• Widget Type: KPI
• Data Source: Main Project Taible
• Metric: Sum of AI Cost / Count of Deployed Features

Why it matters: Usually shows an incredibly low number (e.g., $1.12), which is a powerful metric for management.

---

6. Iteration Efficiency (Line Chart)

Purpose: See if AI agents are getting "smarter" or making fewer mistakes.

Configuration:

• Widget Type: Line Chart
• X-Axis: Deployment Date (Weekly)
• Y-Axis: Average Iteration Count
• Goal: Trend line should go DOWN over time.

Why it matters: High iterations mean prompts need tuning. Low iterations mean agents are nailing it first try.

---

Best Practices for Dashboards

1. Keep it Real-Time: Set refresh rates to 5-10 minutes.
2. Actionable: Every widget should prompt a question ("Why is coverage dropping?").
3. Visible: Put this on a TV screen in your engineering area (or Slack channel).
4. Filterable: Allow filtering by "Epic" or "Priority" to drill down.