In 2026, anyone can open a tool, describe their business and watch a site appear. So the question is no longer « can I do it on my own? » but « how far ». It deserves an honest answer, because the answer is not « nowhere ».

What you have every interest in doing yourself

Writing your content, first of all. Nobody knows your offer better than you, and the exercise forces you to put your ideas in order: what you sell, to whom, in what order it gets told. Doing it in a real page rather than in a Word document changes everything: an overlong title shows, a confused promise jumps out at you. And you can iterate quickly, rephrase, move things around, refine your proposition by watching it exist. That work is not tinkering. It is strategy, and nobody can do it in your place.

What gets paid for later

The risks are not where you expect them. The first: a generic art direction, the consensus of everything the machine has already seen, sometimes even beside the point of your subject. The site looks like a thousand others, it does not look like you. The other risks are invisible in the preview: hosting, updates, the technology that decides whether your back office obeys you or resists you, the SEO that decides whether you get found, the UI and UX choices that decide whether you get understood. None of that shows on launch day. All of it gets paid for afterwards.

The line between what you do yourself and what commits everything that follows A horizontal axis at the top, bounded on the left by « every interest in doing yourself » and on the right by « identity, technical foundation, structure ». A vertical accent stroke, slightly left of centre, cuts the axis and separates two zones. The left zone is a filled panel carrying four right-aligned labels: writing your content; putting your ideas in order; rephrase, move, refine; the best possible brief. The right zone is a dashed frame carrying six left-aligned labels: a generic art direction; hosting; updates; the technology; the SEO; the UI and UX choices. Under the vertical stroke, the words « the choices commit everything that follows ». EVERY INTEREST IN DOING YOURSELF IDENTITY, TECHNICAL FOUNDATION, STRUCTURE writing your content putting your ideas in order rephrase, move, refine the best possible brief a generic art direction hosting updates the technology the SEO the UI and UX choices the choices commit everything that follows
Everything on the left can be replayed at will, everything on the right gets paid for afterwards.

Prompting is not steering

And then there is the wear. When you do not understand what the tool built, the slightest change goes through a prompt: shift a block, prompt; adjust a margin, prompt; three rewordings for a detail a trained hand settles in thirty seconds. The speed you started with turns into dependency.

The cost of the same change along two routes Two rows at the same scale. The top row, named « through a prompt » on the left, lines up three identical rectangles drawn in the accent colour and carries on the right the words « three rewordings ». The bottom row, named « a trained hand », lines up a single rectangle of the same size and carries on the right the words « thirty seconds ». THE SLIGHTEST CHANGE SHIFT A BLOCK, ADJUST A MARGIN through a prompt three rewordings a trained hand thirty seconds
The same detail, and it is that repeated gap that turns the speed you started with into dependency.

So, how far? Up to the line where the choices commit everything that follows: the identity, the technical foundation, the structure. Go at it wholeheartedly, write, structure, iterate, arrive with a draft that says exactly what you want. That is the best possible brief. Then accept the limit of the exercise: knowing how to mix cement and carry bricks is precious on a building site. It does not design the house.