Skip to main content

AI-Assisted Development

AI tooling has shifted from smart autocomplete to full agentic workflows. This section covers how the Aliz web team uses these tools effectively — from writing better prompts to orchestrating multiple AI agents on complex tasks.

note

This space is moving fast. Tool names, features, and best practices in this section reflect the state of the ecosystem in mid-2026. Treat specific feature details as a starting point and verify against the official docs for the tool you're using.

What Does "AI-Assisted" Mean?​

The term covers a wide spectrum:

Autocomplete → Chat → Agents → Multi-agent pipelines

At one end, AI suggests the next line of code. At the other, a network of autonomous agents can plan, implement, test, and document an entire feature with minimal human input.

The important distinction for the Aliz team is between:

  • AI-assisted development — the developer drives. AI accelerates specific tasks: writing boilerplate, explaining code, suggesting fixes. Every output is reviewed before it lands.
  • Vibe coding — the AI drives. The developer describes intent in natural language and largely accepts the output. Useful for prototypes; not appropriate for production. See Vibe Coding.

In both cases, you remain accountable for every line of code that ships.

The Tool Landscape​

IDE Assistants​

  • GitHub Copilot — Microsoft's AI coding assistant for VS Code and JetBrains. Inline completions, chat, multi-file edits, and agent mode. The current team default.
  • Cursor — AI-first IDE built on VS Code. Deep codebase awareness via semantic search; Composer for multi-file editing.
  • Windsurf — Another AI IDE worth knowing; similar positioning to Cursor.

Chat & Reasoning Models​

  • Claude (Anthropic) — Strong long-context reasoning, code review, and architectural discussion. Also the backbone of Claude Code.
  • ChatGPT / GPT-4o (OpenAI) — Versatile; wide plugin and API ecosystem.
  • Gemini (Google) — Large context window; well-integrated with Google Cloud and Workspace.

These models are valuable as a thinking partner, not just a code generator. Ask them to critique a design decision or explain a trade-off.

Coding Agents​

  • ⭐ Copilot Agent Mode — Recommended default. Built into VS Code Copilot Chat. Can read/write workspace files and run terminal commands, with human approval at each step. The lowest-friction starting point for the team.
  • ⭐ Claude Code — Recommended for complex tasks. Anthropic's CLI-based agent. Operates on the full repository with strong multi-step reasoning; best for large cross-file refactors, documentation, and long-horizon tasks. See AI Coding Agents.
  • Other options include Cline (open-source VS Code extension, bring-your-own-key) and Cursor's Composer — worth knowing, but not the team default.

Frameworks & Orchestration​

  • For building custom multi-agent pipelines in Python, OpenClaw is an emerging open-source option (MIT licensed) with a plugin system and pipeline chaining. See Multi-Agent Orchestration for framework comparisons.

Rapid Prototyping Tools​

bolt.new, v0.dev, and lovable.dev are purpose-built for fast, AI-driven prototyping — full apps from a text description, no local setup. See Vibe Coding.

What's in This Section​

  • AI Coding Guidelines — Team-wide rules for using AI tools responsibly: review practices, security hygiene, and when to be careful.
  • Prompt Engineering for Code — Writing effective prompts, workspace instruction files, and tool-specific techniques for Copilot and Cursor.
  • Copilot Customization Files — Scoped instruction files, reusable prompt templates, and agent skills: the file-based system for shaping Copilot behavior at the repo level.
  • Vibe Coding — What it is, which tools enable it, and when it's appropriate vs. when it isn't.
  • AI Coding Agents — How agents work, our recommended agents (Copilot Agent Mode and Claude Code), and practical tips for agent-driven tasks.
  • Multi-Agent Orchestration — Patterns and frameworks for coordinating multiple agents on complex tasks.
  • MCP Servers — Model Context Protocol: what it is, the most useful reference servers, and how to configure it in VS Code and Cursor.
  • Token Consumption Optimization — Practical strategies to reduce token usage and AI tool costs, ordered by impact.
tip

New to AI tooling on the team? Start with Guidelines, then Prompt Engineering.

Further Reading​