How we use AI to build products
Published
Yes, Choie uses AI to write code. It makes implementation faster; it does not change who is responsible for the result. Senior product judgment still directs architecture, testing, review, and quality, and you own every line that ships. This page explains what AI does here, what stays human, and how it affects price and time.
What AI does here
AI coding assistants and agents work inside the editor, directed by the people who own the project. They draft implementation from a decided design, generate the repetitive parts of interfaces and data handling, propose tests, and carry out mechanical refactors across a codebase. This site is built the same way.
What that buys is speed on the parts of software that were already decided. It does not decide anything. AI-accelerated web and mobile apps, directed by senior product judgment. Architecture, testing, and quality stay our responsibility. You own every line.
What stays human
- Deciding what to build. Flows, hierarchy, what ships first, and what is cut. These are product calls made with you, before any code is drafted.
- Architecture. How the product is structured, what it depends on, and how it will hold up next year, not only on launch day.
- Review. Every line that ships is read by a senior person who understands the product. Generated code that is not understood is not merged.
- Testing and quality. Tests, edge cases, accessibility, and performance are checked against real behaviour, then against real people after launch.
- Responsibility. Choie answers for the result. Defects in our own work are fixed at no charge for 30 days after launch.
How it changes price and time
The time saved is spent, not pocketed. It is part of why a coded prototype of one critical workflow fits in a single week for a fixed $4,900, and why a full build of a web or mobile product fits in four to six weeks from $19,000. Fewer hours go into typing; more go into decisions, testing, and watching people use the product. Prices and what they include are on the pricing page; the weekly shape of a build is on the process page.
What it does not change
- You still own the designs and the code once a deliverable is paid for. See who owns the code.
- Third-party libraries and fonts still carry their own licences, and are identified when relevant.
- The written scope, the fixed price, and the live link every Friday are the same with or without AI in the loop.
- Nothing is claimed about your product that we have not seen real people do with it.
What to ask any studio that uses AI
- Who reads every line before it ships, and do they understand the product it belongs to?
- How is generated code tested, and against what?
- Who is responsible for defects, and for how long after launch?
- Do I own the code, including the parts drafted with AI, once I have paid?
- Which third-party components are in the product, and under which licences?
Choie's answers are on this page, the Terms of Use, and the pricing page. Anything unclear, ask directly.
Questions about AI in the work
Does Choie use AI to write the code?
Yes. AI accelerates implementation, while senior product judgment directs architecture, testing, review, and quality. It makes delivery faster; it does not change who is responsible for the result.
Who is responsible if AI-assisted code has a defect?
Choie is, exactly as with any other code. Defects in our own work are fixed at no charge for 30 days after launch. The tool used to write a line does not change who answers for it.
Do I own code that was written with AI assistance?
Yes. Everything delivered is a deliverable under the same terms: once its fee is paid in full, the designs and source code are yours to use, modify, and distribute without restriction.
Does AI make the product cheaper or faster?
Faster, and it is part of why a coded prototype fits in one week and a full build in four to six weeks at fixed prices. The time saved goes into decisions, testing, and watching real people use the product rather than into typing.