The last lesson built one AI feature in detail. This lesson zooms out to where AI fits across a whole web development workflow, beyond writing individual lines of code.
Beyond line-by-line coding help, some tools generate a working starter site or app from a plain-language description — a homepage layout, a basic app structure, a component library already wired together. These are genuinely useful for skipping the blank-page problem, but the output is a starting point to build on and understand, not a finished, production-ready product to ship untouched.
The same assistant that helps with HTML and CSS is just as useful for backend logic — describing a data operation in plain language and getting working server-side code back, whether that's a simple form handler or a full CRUD (create, read, update, delete) flow. The same caution from the coding assistants lesson applies doubly here: backend code touches real data, so review and test it thoroughly rather than trusting it by default.
A REST API is a standard way for one piece of software to request data or trigger an action in another over the internet — it's exactly how the chatbot in the last lesson talked to an AI model, and it's the same underlying pattern used to connect a website to almost any external service: payment processing, maps, weather data, or another AI provider entirely.
AI tools can help with both sides of search engine optimisation: the technical side (flagging missing meta descriptions, broken heading structure, slow-loading pages) and the content side (drafting page copy, suggesting keyword-relevant headings). As with any AI-generated content, technical SEO suggestions are worth verifying against current best practice, and written content still needs a human review pass before publishing.
Writing a clear description of a set of code changes — a commit message — is a small but constant task, and AI is genuinely good at drafting one from the actual changes made. It won't know the full context of *why* a change was made unless you tell it, so a quick human edit still matters, but it removes the friction of staring at a blank commit-message box.
A site with an AI feature has one extra requirement beyond ordinary static hosting: somewhere to run the small server-side proxy script that keeps the API key safe, as covered in the last lesson. Most modern hosting platforms support this directly, so it rarely means running your own physical server — but it does mean "just static file hosting" alone isn't enough once an AI feature is involved.
The Pattern Repeats Here Too
Every one of these use cases follows the same shape as the individual coding-assistant lesson: AI removes friction and speeds up a task you already understand, rather than replacing the need to understand it. That distinction is what separates using AI well from shipping something broken faster.
Web development is one path AI accelerates. The next two lessons cover a different one — using Python to work with AI programmatically, starting from the very basics.