Design-led AI MVP · from €4,500 · 2 weeks
An idea is cheap to argue about.
Build it and find out.
Two weeks. Research, information architecture, interface and a working product that is actually deployed. One person the whole way through, so nothing is lost handing the design to whoever builds it.
Who it's for
-
You want to know if the idea works before committing engineering.
A working product answers that in two weeks. A specification argues about it for two months.
-
You have an internal concept that needs to be shown, not described.
Slides get nodded at. Something people can click gets a decision.
-
You are pitching, and a prototype is not enough.
Investors and steering committees can tell the difference between a demo and a product.
The work
Concept and build, same engagement.
The usual split puts research in one place and implementation in another. Most of what gets lost, gets lost in between.
- 01
Lean research
Who this is for, what they are trying to do, and what already exists. Short, but not skipped: everything after this inherits it.
- 02
Flows and information architecture
The structure, before any screen. This is the cheapest place to be wrong and the most expensive place to stay wrong.
- 03
Interface on a real design system
Tokens, components and states, not a set of one-off screens. It is what makes the thing extensible after I leave.
- 04
Built and deployed
I build it with AI assistance and put it online. Same person throughout, so nothing is lost explaining the design to whoever implements it.
What you get
A product, not a deck.
Everything is yours at the end: the running product, the code, and the system it was built on. Nothing is locked to me.
- A working product, deployed and reachable at a URL
- The complete source code, in your repository
- The design system it is built on, extensible after the engagement
- Concept documentation: the structure, the decisions and what was rejected
- A walkthrough call, and a written list of what I would do next
The two weeks
Ten working days.
-
Day 1
Kickoff
Your goal, your users, the one thing this has to prove.
-
Day 2–3
Research and structure
Context, flows, information architecture.
-
Day 4–6
Interface
Design system and screens, reviewed with you.
-
Day 7–9
Build
Implemented, deployed, tested against the flows.
-
Day 10
Handover
Walkthrough, code, documentation, next steps.
The obvious question
Why not just use Lovable or v0?
For a prototype, do. Those tools are genuinely good, and if the goal is to show a shape to three colleagues on Friday, they will get you there faster than I will.
The difference shows up when the thing has to survive contact with real users, a security review, or a second developer. CodeRabbit analysed 470 open-source pull requests and found AI-generated ones carried roughly 1.7 times more issues than human-written ones, 10.83 against 6.45 on average. That is not an argument against building with AI. I build with it too, and it is why two weeks is enough. It is an argument for someone being accountable for what ships.
What you are buying is not the typing. It is the decision about what to build, the structure underneath it, and someone who has shipped interfaces where being wrong was expensive.
Price
from €4,500
For the two-week engagement: one product, one main user, one thing it has to prove. Anything larger I quote after kickoff, once I know what it actually needs. Rates are ranges, stated plainly.
Most of these do not stop at two weeks. They become product design or an embedded engagement, and I would rather say that up front than discover it in week three.
Questions
Before you ask.
Sixty minutes for kickoff, access to anyone who talks to your users, and one clear answer to what this has to prove. That is enough.
A real one: deployed, with source code you own. It is deliberately small in scope, not deliberately fragile.
From €4,500 for the two-week engagement. Larger scope is quoted after kickoff, once I know what the product actually needs.
That is the point. You get the code and the design system, and the structure is documented so another developer can pick it up.
Always, before we talk specifics. Most of my client work is under one.
Tell me what you want to prove.
Two or three sentences about the idea and what it has to show. I reply within two working days.