Help every team at Volue find the right level of AI-assisted development tooling. Pick the level that matches your role and the complexity of what you're building.
4
Levels
6+
Tools Covered
0 min
To Get Started
🚗
Think of these levels like gears in a car
You don't drive in 4th gear through a parking lot, and you don't stay in 1st on the highway. Most of your daily work will happen in Level 1 or Level 2. You shift up when the task demands it. You don't need to master all levels — start where you are, get comfortable, then explore the next level when you hit a wall.
The Four Levels at a Glance
Level 0
Level 1
Level 2
Level 3
Name
AI Assistants
Developer Fundamentals
Agents & Skills
Full Lifecycle
Tools
Claude.ai, M365 Copilot
Claude Code, GitHub Copilot
Superpowers, Subagents
GSD, BMAD
Who
Everyone
All developers
Developers wanting more
Complex projects
Effort to start
Zero
Minimal
~30 min setup
~1 hour setup + learning
Best for
Writing, research, Q&A
Everyday coding
Specialised or repetitive tasks
Multi-phase feature builds
Governance Tier
Tier 1 — Individual
Tier 1 — Individual
Tier 1–2
Tier 2 — Team
Governance note: All tools referenced in this guide are pre-approved under the AI Governance Framework. Claude is one of Volue's three strategic AI vendor partners (alongside Microsoft and Anthropic). If you want to introduce a tool not listed here, follow the governance framework's tool intake process.
Level 0 — AI Assistants
Everyone
Claude.ai & Microsoft Copilot
For anyone at Volue, regardless of role. No terminal, no code required.
Who should use this: Customer Success, Consultants, Product Managers, Leadership — anyone who doesn't write code daily. Most people will use both: Copilot for Microsoft 365 tasks, Claude for everything else.
Getting Started
Claude.ai: Log into claude.ai with your Volue account. That's it.
Microsoft Copilot: Already available — look for the Copilot icon in Word, Excel, Teams, and Outlook.
Both tools operate under Volue's enterprise agreements and do not use your inputs for model training. Never enter customer PII or classified data — see the Governance Framework's security guidelines for data classification rules.
Level 1 — Developer Fundamentals
All Developers
Claude Code & GitHub Copilot
The foundation. Every developer starts here with two complementary tools.
GitHub Copilot — IDE pair programmer
Real-time code completion as you type — suggestions appear inline
Works inside VS Code, JetBrains, and other supported IDEs
Great for boilerplate, repetitive patterns, and "finish my thought" moments
Copilot Chat for quick questions without leaving your editor
Claude Code — terminal coding partner
Lives in your terminal, understands your full codebase
Writes, edits, and refactors code across multiple files
Runs tests, debugs errors, explains complex logic
Works with git — commits, branches, PRs
How they work together
GitHub Copilot handles the small, fast stuff — autocomplete, single-function generation, inline suggestions while you're typing. Claude Code handles the bigger picture — multi-file changes, debugging sessions, architecture discussions, code reviews. Most developers will have both running: Copilot in the IDE, Claude Code in a terminal alongside it.
Why start here first
Before adding any frameworks on top, get comfortable with the basics. Learn how both tools think, where they're strong, where they need guidance. For Claude Code, a well-written CLAUDE.md file in your repository goes a long way — it tells Claude about your project conventions, tech stack, and patterns.
When this is enough
Bug fixes, small features, code reviews, explaining unfamiliar code, writing tests for existing code, quick prototypes.
When to shift up
You find yourself repeating the same instructions across sessions. You're building something that spans many files and multiple days. You wish Claude was more opinionated about how to approach the work.
Getting Started
GitHub Copilot: Already available — enable the extension in your IDE and sign in with your Volue GitHub account.
Claude Code: Install Claude Code → open your terminal in a project → start talking.
Both tools send code to external APIs under Volue's enterprise agreements. Do not use them on repositories containing customer secrets, API keys, or credentials outside of encrypted vaults. If in doubt, check the security guidelines.
Level 2 — Agents, Skills & Structured Development
Power Users
Teach Claude your team's way of working
This level adds two things on top of vanilla Claude Code: specialist agents and a methodology framework.
Agent & Skill Libraries
Pre-built specialists that Claude can call on. Think of them as expert colleagues sitting next to Claude — a Kubernetes specialist, a security reviewer, a .NET architect. You install the ones relevant to your stack, and Claude automatically delegates to them when appropriate.
Example Agents
A code reviewer subagent that knows Volue's coding conventions
A platform deployer skill that understands our GitOps/Flux workflow
Language-specific agents for .NET, Python, React, or whatever your team uses
Public libraries like VoltAgent with 100+ ready-made subagents
Superpowers (Methodology Framework)
Superpowers changes how Claude works, not just what it knows. Once installed, it enforces professional development practices automatically:
💭
Brainstorming First
Claude asks questions and explores alternatives before writing a single line of code
🧪
Test-Driven Development
Tests are written first, always. If Claude writes code before tests, Superpowers deletes it and starts over
🔍
Systematic Debugging
No more random fixes. A 4-phase root cause investigation process
✅
Code Review
Automated review against the plan after each task is completed
When to use this level
Mid-size features, new services, refactoring projects, anything where "just start coding" leads to rework. Superpowers is especially valuable for developers who want guardrails that prevent shortcuts.
Getting Started
Subagents: Browse available agents, install what fits your stack. See the internal guide for Volue-specific agents.
Restart Claude Code. Done. It activates automatically.
Pre-approved tools: The agents, skills, and Superpowers framework referenced here have been reviewed and pre-approved under the AI Governance Framework. If you discover a new third-party agent or framework, submit it through the tool intake process — the Team Leads evaluate new tools bi-weekly.
Level 3 — Full Lifecycle Management
Complex Projects
For multi-phase projects where structure is everything
When you're building something that takes weeks, involves multiple components, and needs traceability — Levels 1 and 2 aren't enough. Two frameworks serve this purpose. They solve the same problem differently.
GSD (Get Shit Done)
"The complexity is in the system, not in your workflow."
GSD is built around one powerful idea: every task gets a fresh AI brain. Instead of one long session where Claude gradually loses context, GSD spawns a clean instance for each task with a full 200K-token context window.
How it works
/gsd:new-project — GSD interviews you about what you're building
Creates a structured spec, breaks it into phases and atomic plans
Each plan runs in a fresh subagent — your main session stays clean
Independent tasks run in parallel (waves), dependent ones wait
/gsd:verify-work — guides manual testing
Every task gets its own git commit — fully revertable
Best for: Developers who want maximum automation with minimal process.
npx get-shit-done-cc
BMAD
"A full AI development team, with roles and accountability."
BMAD gives you 21 specialised AI agents that map to real development roles — Product Manager, Architect, Scrum Master, Developer, QA. Each has defined responsibilities and workflows.
How it works
Analyst gathers requirements and context
Product Manager creates a detailed PRD
Architect designs the system and makes tech decisions
Scrum Master breaks work into epics and story files
Developer implements each story with full context
QA agent validates against requirements
Best for: Teams that think in agile terms. Projects needing documentation and audit trails.
npx bmad-method install
Which One Should I Pick?
GSD
BMAD
Feels like
A smart project manager that runs your AI workers
A full agile team in your terminal
Context management
Automatic — fresh AI instances per task
Manual — fresh sessions per workflow
Strengths
Speed, parallel execution, zero context rot
Process rigor, documentation, role-based workflows
Learning curve
Lower — fewer commands
Higher — more concepts, very guided
Documentation
Specs and git history
PRDs, architecture docs, story files, sprint tracking
Our recommendation: If you're unsure, start with GSD. Its automated context management means fewer things can go wrong, and you can be productive within minutes. Move to BMAD if you need heavier process structure or your team already works in agile sprints.
Governance note: Full lifecycle frameworks involve team-level tooling decisions and generate project artifacts. These fall under Governance Tier 2 (Team — Guided). They are pre-approved, but if you're using them on a project that spans multiple teams, flag it at the bi-weekly Team Leads for visibility.
Quick Decision Guide
What are you trying to do?
📝
I don't write code
AI Assistants
Level 0
🔧
Quick fix, small task, exploring code
Developer Fundamentals
Level 1
🎯
Building a feature, want quality guardrails
Superpowers + Agents
Level 2
🚀
Multi-week project, new service, major migration
GSD or BMAD
Level 3
Not sure? Start at Level 1. You'll know when you need more.
Champions & Support
🏆
Become a Claude Champion
This guide is supported by Volue's AI Champions Program (Creators & Ambassadors). Champions are recognised through contribution, not appointment — if you actively help others adopt these tools, you're building your path to Champion and Distinguished status.
🛠
Creators Track
Build Claude-specific prompts, agents, skills, or CLAUDE.md templates and contribute them to the AI Catalogue. Every asset you build is automatically tracked.
📢
Ambassadors Track
Train colleagues on Claude Code, present tips at Knowledge Sessions, help others troubleshoot at Office Hours. Spread adoption across your team and BU.
Get Help & Share What You Learn
💬
AI Office Hours
Bi-weekly
Drop in with questions, get live help, demo your setup
