Bringing work context to your code in GitHub Copilot

The evolution of AI-driven developer tools has largely focused on accelerating code generation and reducing repetitive tasks. However, as I see it, the next significant leap is not about writing code faster but enriching the coding experience with broader organisational knowledge. The recent introduction of the GitHub Copilot SDK and its integration with Work IQ, the intelligence layer behind Microsoft 365 Copilot, marks a noteworthy shift in this direction. This development aims to bridge the persistent gap between code and the context that surrounds it—requirements, ownership, design intent, and organisational memory.

Reframing Developer Productivity with Contextual Intelligence

In my experience, developers rarely struggle with just writing code. The real challenge often lies in tracing design decisions, understanding why something was built a certain way, and identifying who owns or understands a specific part of a system. The new capabilities demonstrated by GitHub Copilot CLI and its integration with Work IQ attempt to surface this hidden but critical context directly in developers’ workflows.

The original article outlines several practical experiments where contextual awareness transforms routine developer activities. These examples highlight both tactical improvements in productivity and strategic opportunities for organisations seeking to reduce friction in their software delivery processes.

Finding the Right Owner for a Piece of Code

Ownership discovery is an age-old pain point, especially as teams grow and contributors cycle through projects. Traditional methods like `git blame` often return outdated or unhelpful results—sometimes surfacing names no longer associated with the project.

With Copilot now able to aggregate signals from commit history, project metadata, meetings, emails, and internal documents, developers can identify current owners quickly without interrupting colleagues or sifting through multiple systems. This capability relies on connecting GitHub Copilot to Work IQ’s organisational knowledge graph.

Strategic implication: For technology leaders managing large or distributed teams, this feature has potential to reduce onboarding time for new staff and mitigate knowledge silos that slow down incident response or project handovers.

Creating an Architecture Diagram from a Meeting Transcript

Translating spoken discussions into technical artefacts remains error-prone and time-consuming. The integration between Work IQ and Copilot CLI allows meeting transcripts to be parsed for architectural components and relationships, enabling automatic generation of draft architecture diagrams.

A video demonstration in the article (see example) shows Copilot extracting relevant details from transcripts and synthesising them into visual representations—a task that previously required manual effort across multiple tools.

My analysis: Automating this process not only saves time but also ensures alignment between design intent discussed in meetings and technical implementation captured later. For leaders overseeing complex projects or compliance-heavy environments, this reduces risk associated with miscommunication or drift from agreed designs.

Comparing an Implementation to the Original Design Spec

One persistent issue in software development is divergence between initial design specifications and final implementations. The enhanced Copilot can now reference relevant design documents directly alongside source code changes to highlight areas where requirements have changed or were not met.

Instead of relying on manual reviews or tribal knowledge to spot discrepancies, this approach provides immediate visibility into gaps between intended architecture and actual deployment.

Business value: For organisations operating under regulatory scrutiny or strict quality controls, this feature adds defensibility to their change management process while freeing up senior engineers from routine audit tasks.

Bringing Work Context into Your Own Apps

Perhaps most impactful is the release of the GitHub Copilot SDK (currently available in Technical Preview). With just a few lines of code, teams can embed contextual intelligence agents into their own applications—tailoring how work context appears within custom workflows.

An example provided involves a lightweight Visual Studio Code extension surfacing meeting notes, design docs, and files from sources like SharePoint or OneDrive directly inside the editor window (see demo). This reduces tool-switching overheads that often break developer flow.

Strategic perspective: As more business-critical information resides outside traditional code repositories—in chat logs, document stores, or email—equipping development environments with seamless access to these resources will become essential for productivity at scale.

Technical Enablement: Getting Started with Work IQ Integration

To leverage these features today:

• A GitHub Copilot subscription is required
• Access to Microsoft 365 subscription including Microsoft 365 Copilot is necessary
• Tenant admin approval must be secured (details are available via the Work IQ MCP Server Repo)
• Installation steps:
• Ensure you have the latest version of GitHub Copilot CLI
• Use `/plugin marketplace add github/copilot-plugins`
• Install using `/plugin install workiq@copilot-plugins`
• Restart the CLI to activate Work IQ MCP server

The article notes that eligible users may obtain a free M365 dev tenant through Visual Studio subscriptions or via the Microsoft 365 Developer Program.

Strategic Recommendations for Technology Leaders

As I reflect on these advancements, several actionable insights emerge for CIOs, engineering directors, and heads of developer experience:

• Assess Knowledge Fragmentation Risks: Evaluate where critical contextual information lives today—if much resides outside your source control systems (in documents or emails), consider pilot deployments of context-enriched tools.
• Strengthen Compliance Audits: Use automated comparison between code implementations and original specifications as an audit trail for regulated industries.
• Accelerate Onboarding: Leverage ownership mapping features within large teams to flatten learning curves for new hires or rotating staff.
• Promote Flow Efficiency: Reduce cognitive load by bringing supplementary documents directly into IDEs rather than expecting developers to switch contexts repeatedly.
• Experiment Responsibly: While early-access features like the GitHub Copilot SDK offer promise, ensure robust governance around data privacy when connecting organisational knowledge graphs across platforms.
• Monitor Productivity Metrics: Track whether tool adoption leads to measurable reductions in time spent searching for documentation or clarifying requirements post-handoff.
• Encourage Customisation: Explore how embedding contextual agents via SDK can address team-specific workflows rather than relying solely on out-of-the-box integrations.

Conclusion: From Code Completion to Knowledge Completion

The integration of organisational context into developer tooling represents more than incremental productivity gains—it signals a maturation in how we think about AI assistants in professional environments. By moving beyond isolated code suggestions towards holistic work understanding powered by platforms like Work IQ and GitHub Copilot SDKs, technology leaders have an opportunity to create genuinely intelligent development ecosystems.

In my view, those who invest early in aligning their workflows around these new capabilities will achieve not just faster delivery but greater resilience against knowledge loss and operational bottlenecks—outcomes that matter deeply as software complexity grows.

For those interested in exploring these ideas further or experimenting with contextual plugins themselves, I recommend reviewing Kayla Cinnamon’s original blog post: Bringing work context to your code in GitHub Copilot.

Want more cloud insights? Listen to Cloudy with a Chance of Insights podcast: Spotify | YouTube | Apple Podcasts

---

Source: https://devblogs.microsoft.com/blog/bringing-work-context-to-your-code-in-github-copilot