GitHub outlines agentic engineering workflow to address 441% increase in code review times

Details

GitHub reports that while commit volume surged over 400% in one week in October, a Faros AI study found pull requests spent 441% longer in review as teams shifted from low to high AI adoption.
The article defines an "agentic engineering" workflow across four stages: Plan, Build, Verify, and Observe, designed to help developers manage context, execution, review, and post-deployment issues.
In the Plan stage, Copilot integrates context from Slack, Teams, and connected tools to generate implementation plans and GitHub Issues, using Agentic Workflows for triage and Plan Mode for complex decisions.
The Build stage utilizes session continuity, cloud sandboxes, and isolated repository copies to allow parallel task execution; Copilot Fusion allows users to select models and reasoning depth, while Dreaming maintains shared memory and instructions.
Verification relies on stacked pull requests, GitHub Code Quality checks, and AI Scan for vulnerabilities beyond CodeQL coverage; Agent Merge and Agentic Autofix handle routine fixes and merge conflicts.
The Observe stage connects Copilot to observability tools via connectors like Fabric to investigate production issues, draft issues with regression tests, and refine future guidance based on operational evidence.

Key point