Glass Factory v2.5: AI-Powered Sprint Intelligence

Release Overview

We are excited to announce Glass Factory v2.5, our most significant platform update of the year. This release introduces AI-powered sprint intelligence capabilities that help engineering teams plan more accurately, identify risks earlier, and deliver more predictably.

Key Features

AI-Powered Velocity Prediction

Glass Factory now uses machine learning models trained on your team historical data to predict sprint velocity with unprecedented accuracy.

How It Works:

Input Signals:
├── Historical velocity (weighted recency)
├── Team composition changes
├── Holiday and PTO calendar
├── Story point distribution patterns
├── Technical debt ratio
└── External dependency status

Output:
├── Predicted velocity range (P25, P50, P75)
├── Confidence interval
├── Risk factors identified
└── Recommendations for scope adjustment

Accuracy Improvements:

Automated Capacity Planning

The new Capacity Planner automatically calculates team availability based on:

• Connected calendar integrations (Google Calendar, Outlook)
• PTO requests submitted through Glass Factory or HRIS integrations
• Historical meeting load patterns
• On-call rotation schedules

Intelligent Dependency Detection

Glass Factory v2.5 automatically identifies cross-team dependencies by analyzing:

1. Code references: Detecting imports/calls to other team codebases
2. API contracts: Identifying shared service dependencies
3. Historical patterns: Learning from past dependency-related delays
4. Explicit links: User-defined dependency relationships

Migration Guide

Breaking Changes

v2.5 includes the following breaking changes:

1. API Version: The /api/v1/sprints endpoint is deprecated. Use /api/v2/sprints
2. Webhook Payloads: Sprint event webhooks now include aiInsights field
3. Permission Model: New sprint.ai.view and sprint.ai.configure permissions

Upgrade Steps

# 1. Update Glass Factory CLI
npm install -g @uranuslab/glass-factory-cli@2.5

# 2. Run migration script
gf migrate --from 2.4 --to 2.5

# 3. Update API integrations
gf api:check --compatibility v2

# 4. Configure AI features
gf ai:setup --training-window 90d

Configuration Options

New configuration options in glass-factory.config.js:

module.exports = {
  ai: {
    enabled: true,
    velocityPrediction: {
      confidenceThreshold: 0.75,
      trainingWindow: "90d",
      excludeOutliers: true
    },
    dependencyDetection: {
      codeAnalysis: true,
      apiContractScanning: true,
      sensitivityLevel: "medium"
    },
    capacityPlanning: {
      calendarIntegration: "google",
      bufferPercentage: 15
    }
  }
};

Availability

Glass Factory v2.5 is available now for all Enterprise customers. Professional plan customers will receive access on March 1st, 2026.

Getting Started

To enable AI-powered sprint intelligence:

1. Navigate to Settings → AI Features
2. Enable "Sprint Intelligence"
3. Configure training data window (recommend 90 days minimum)
4. Review and approve AI model access to historical data

For questions or assistance with migration, contact your Customer Success Manager or reach out to support@uranuslab.com.