Define the North Star metric for Copilot serving developers

Problem Statement Description

Product context: Microsoft is a productivity, software, AI, gaming, and cloud company; its products include Windows, Microsoft 365, Teams, LinkedIn, Xbox, Azure, Dynamics, and Copilot.

Microsoft Copilot for developers sits inside the day-to-day software creation workflow: understanding a task, writing or editing code, getting suggestions in an IDE, reviewing generated changes, running tests, debugging, and preparing work for merge. The product is used by individual developers, teams, and enterprises that care about productivity, code quality, security, compliance, and developer trust.

Your task is to define a North Star metric for this product. The metric should capture whether Copilot is creating durable value for developers, not just whether it is producing AI outputs or driving superficial usage. Consider how developers actually adopt AI assistance across coding sessions, repositories, languages, seniority levels, and enterprise environments.

A strong answer should make clear what behavior is being measured, who is included, over what time window, and why the metric is useful for product decision-making. It should also address how Microsoft would instrument the metric across IDEs, repositories, developer accounts, and enterprise tenants while respecting privacy and security expectations.

The experience should consider:

- The core developer workflow where Copilot creates value, from suggestion to accepted code to shipped work.

- A precise metric definition, including numerator, denominator, time window, and eligible user population.

- How to distinguish meaningful developer productivity from raw activity, questions, completions, or engagement volume.

- Cohorts such as individual developers, enterprise teams, new vs. retained users, language/framework, IDE, repository type, and experience level.

- Instrumentation needs across editor events, acceptance/editing behavior, build/test outcomes, pull requests, and enterprise admin contexts.

- Guardrail metrics for code quality, security, hallucination risk, developer satisfaction, latency, cost to serve, and compliance.

- How the metric would help Microsoft make roadmap, go-to-market, pricing, and platform integration decisions.

- Risks of gaming or misinterpreting the metric, especially in complex team-based software development environments.

The goal is to propose a North Star metric that reflects sustained developer value for Copilot, can be measured responsibly at Microsoft scale, and gives product leaders a clear signal for whether the product is improving the way developers build software.