Web Directions Year round learning for product, design and engineering professionals

A pragmatic guide to how AI is impacting the practice of front‑end development: agents, MCP, and how to thread them into the work you already do.

AI is edging into the daily work of front‑end teams—component libraries, design reviews, code review, testing, even the architecture decisions that shape our products. So, our Developer Summit is designed to separate signal from noise.

Why AI matters to front‑end

Front‑end used to mean “the UI.” Today it’s the UI and the workflows that produce it. Agents and the Model Context Protocol (MCP) are reshaping those workflows:

• Agentic code assistance: bounded tasks (writing tests, migrating components, refactoring styles) are increasingly delegated to capable agents—with review gates you control.
• MCP as the plumbing: instead of brittle ad‑hoc automation, MCP servers give you a standard way to connect tools—design tokens, repos, issue trackers, CI.
• Design→Code loops: assistants that read design system docs, Figma files, and your component library to propose changes that match house style.

The win is not “AI writes the app.” It’s that your practice speeds up while staying consistent with your standards.

Practical adoption patterns

1. Automate what you already standardised
   Linting rules, component scaffolds, test templates, docs conventions—these are ideal for agent workflows. Start where you can easily evaluate correctness.
2. Wrap risky tasks in reviewable specs
   Treat agent tasks like PRs: clear specs in, diffs out, humans approve. MCP helps here by brokering context safely.
3. Instrument the pipeline
   Track time saved on routine work (tests, stories, docs), not “AI lines of code.” Your KPI is cycle time and defect rate.
4. Plan for offline and privacy
   For regulated work or sensitive assets, evaluate on‑device and on‑prem options; pair with the browser’s offline capabilities (Service Workers, OPFS) for field use.

What you’ll learn at Developer Summit

• Agents you can reason about: small, auditable agents in JavaScript—where they help, where they get in the way.
• MCP in the toolchain: connecting design tokens, repos and CI without glue‑code sprawl.
• Human‑in‑the‑loop patterns: design, review and governance so teams stay accountable.

Register for Web Directions Developer Summit and join us November 19-20 in Sydney (or streaming online) to explore these transformations with the people driving them.