2026
AsMade
Showing the AI-assisted process, not only the result
Case study by Ian Kirs · ARCHKIRS
- Research Project
- Web Development

About the project
AsMade is a research-stage web project about transparency in AI-assisted work. Its goal is to show not only the final output but the path behind it: which materials a person used, what actions they took, which decisions they made, and where AI became part of the process.
The central concept is the MADE Record, a concise structured account of Materials, Actions, Decisions, and Evolution. It is designed as a clear way to explain how work was made without trying to calculate an automatic “AI percentage.”
The current public site is a system of audience-specific landing pages that leads participants into a pilot flow. AsMade is not presented as an AI detector, legal certification, or proof of authorship.
Challenge
Explain a complex and sensitive topic - transparency in AI-assisted work - without making the project look like another AI detector, while keeping it understandable for students, creators, and reviewers.
Solution
Instead of one generic landing page, the site uses a small system of connected pages: general, student, artist, and reviewer. Each explains the same product through the needs of a specific audience while preserving the shared terms AsMade, MADE Record, and MADE. Pilot applications use one shared form with no file upload at the application stage.
What was implemented
- 01
Four audience flows: general, student, artist, and reviewer, plus compatibility redirects for earlier route names.
- 02
Consistent product naming: AsMade for the project, MADE Record for the result, and MADE for Materials, Actions, Decisions, Evolution.
- 03
Audience-specific value propositions for students, creative practitioners, and reviewers without changing the core product idea.
- 04
Shared Tally pilot flow: application first, then work with a real non-confidential project instead of collecting sensitive files on the first screen.
- 05
Explicit claim boundaries: the site does not promise AI detection, legal certification, proof of authorship, or replacement of human judgment.
Gallery
Business value
- A complex idea is explained through concrete user scenarios instead of one overloaded technical description.
- One product architecture supports several audiences and makes it possible to test which framing works best for each.
- Careful claims support trust: the product explains what it does and is equally clear about what it does not do.
- The landing system already supports pilot recruitment and provides a foundation for developing the full MVP later.
Need a similar system?
If you need more than a website - a connected work tool, client portal, or automation for a specific process - tell me about the task. We can start with a focused part and expand the system gradually.