Every new tool promises to change how websites are made. Most of them change one step. AI is different because it touches all of them at once: research, design, code, content, testing and the work that happens after launch. Here is what that looks like inside a real project, and what it means if you are about to commission a website.
Research takes days, not weeks
Discovery used to mean reading every competitor site, every review and every support ticket by hand. Now AI does the first pass. It groups hundreds of customer comments into themes, maps the words people actually use and drafts a competitor overview in an afternoon.
The judgement stays with us. A model can tell you that customers mention price forty times, but it cannot tell you that the real problem is trust. We spend the time we save talking to more real people, not fewer.
Design explores more directions
Early in a project we now explore far more routes than we used to. AI helps us sketch layout variations, rough imagery and copy drafts in minutes, so a workshop can react to twelve ideas instead of three.
What it does not do is decide. Taste, brand fit and the small details that make a site feel premium still come from designers who know the client and the audience. AI widens the funnel. People choose what goes through it.
Code ships faster, and cleaner
On the development side, AI pair programming has changed the rhythm of a build. Components, tests and the repetitive parts that used to take a day are drafted in an hour, then reviewed line by line by a developer. The time we save goes into the parts that make a site memorable: motion, interaction and performance.
Where it helps most
Writing and updating tests
Moving content between systems
Turning designs into first drafts of components
Catching accessibility issues before a human review
Where we keep it on a short leash
Anything that touches security, payments or personal data. Generated code is a draft, never a deliverable, and nothing ships without a review and a test.
Content is written for people and for machines
Search is changing too. More visitors now meet a brand through an AI answer before they ever see its website. That means content has to be clear enough for ChatGPT, Perplexity and Google’s AI Overviews to quote, and structured so they understand who you are and what you offer.
We call this answer engine optimisation, and it now sits next to classic SEO in every build. AI helps us test how a page is summarised and where it gets cited, so we can close the gaps before launch.
Sites keep improving after launch
The work no longer stops at go-live. AI assistants answer visitor questions, route new leads to the right person and flag the pages where people drop off. Small automations keep content fresh, and reports arrive without anyone building a spreadsheet.
What does not change
AI is a powerful tool, but a website is still a set of decisions: what to say, who to say it to and how it should feel. Those decisions need people who understand your business and care about the result.
For our clients the change is simple. Projects move faster, budgets go further and more of the time goes into the craft that makes a site worth remembering. If you are planning a new website and want to see how this works in practice, book a free call with the studio.
