Workshops: Agentic Coding

GitHub Copilot Workshop: Agent Mode, Code Review & Rollout

Many teams pay for GitHub Copilot but use little beyond autocomplete. In this workshop developers learn to use Agent Mode, the Copilot coding agent and Copilot code review productively, and engineering leads get a plan for rolling it out across the team.

Who is this workshop for?

  • Development teams on Copilot Business or Enterprise
  • Tech leads and engineering managers
  • Platform and developer experience teams
  • GitHub organisation admins
  • Teams moving from autocomplete to agentic work

What you will take away

  • Use Agent Mode in VS Code or JetBrains for multi-step tasks while staying in control.
  • Delegate well-scoped issues to the Copilot coding agent and review its pull requests efficiently.
  • Use Copilot code review as an extra check without replacing human review.
  • Maintain repository instructions and prompt files so Copilot follows your conventions.
  • Leave with a rollout plan covering policies, model choice and success metrics.

Agenda

  1. Module 1

    From suggestions to agents

    • Completions, chat, edits and Agent Mode compared
    • Model selection in Copilot and when each is worth it
    • Live demo on your repository
  2. Module 2

    Agent Mode in practice

    • Exercise: build a multi-file feature with tests
    • Steering context: files, terminal, MCP servers
    • Common failure modes and how to recover
  3. Module 3

    Coding agent and code review

    • Writing issues the coding agent can actually solve
    • Reviewing and iterating on agent pull requests
    • Configuring Copilot code review and setting expectations
  4. Module 4

    Team conventions

    • copilot-instructions.md and prompt files
    • Shared prompts for recurring tasks
    • Brief comparison with Cursor, Claude Code and OpenAI Codex
  5. Module 5

    Rollout and governance

    • Policies, content exclusions and public code filter
    • Seats, usage analytics and a champions programme
    • Metrics that actually mean something

Why so many Copilot seats go underused

Copilot has grown from an autocomplete tool into an agent platform. Many teams are unaware of the newer features or had poor first experiences because tasks were too vague.

The workshop shows on your own code when Agent Mode and the coding agent add real value, and which tasks are better left to humans.

Security and governance

For organisations, settings at the org and enterprise level are decisive. We walk through the relevant policies with your admins.

  • Content exclusions for sensitive files
  • Filtering suggestions matching public code
  • Enabling models and preview features
  • Handling MCP servers and external tools

Your trainers

The workshop is run by neooutput AI Code Engineers who use GitHub Copilot and other agents daily in client projects. Optional follow-up coaching after the first weeks is available.

Frequently asked questions

Is this worth it if we have used Copilot for a year?

Especially then. Most teams use only a fraction of the features. We start with a short assessment and focus where the leverage is greatest.

Do we work with our own code?

Yes, where possible. Alternatively we provide a realistic practice repository similar to your stack.

Does Copilot code review replace our reviews?

No. It is a useful first pass that catches routine issues. Architecture, domain logic and accountability stay with the team.

Workshop

Get in touch

Shall we turn your Copilot seats into real productivity?

Tell us a little about your team – we will get back to you within 24 hours with a proposal for content, format and dates.

Prefer to talk?

We only use your details to handle this request. See our privacy policy.