Clickable prototypes for companies in Germany

What a clickable prototype decides

A clickable prototype lets people decide on software by using it before it exists. Management sees the product it would pay for, and future users show where they would get lost.

A software decision often rests on a written concept and a few slides. Each reader imagines a different product, and the differences show up once the build has started. A prototype puts one version of the product in front of everyone who decides.

How I build a prototype for a decision

I start with the flow a user goes through, from the first screen to the end of the task. The screens get real words from your business and example data, because placeholder text hides the questions users would ask.

Where the user works shapes the screens. A service technician who opens the app on a phone at a customer’s site needs other screens than a clerk at a desk with two monitors, so I ask about that situation before I draw the first screen. The labels name the next step: a button that says “To the offer” tells the user what happens, and “Next” leaves it open. Details: Interaction design.

The test with users is short. I give each user a task, such as finding a price or sending a request, and watch without helping. Where several users stop at the same screen, that screen changes before the decision.

Prototype, proof of concept or MVP

Three ways to test a software idea before the full build. Each answers a different question, and the question decides which one your company needs.

	Clickable prototype	AI proof of concept	MVP
The question it answers	Do users understand the flow, and do we want this product?	Does the AI do this task on our data well enough?	Do real users use the product for their job?
What it is	Screens that react like the app, with example data	A working test of one AI task, measured by a script	Working software with a database
Who uses it	Management and a few future users in a test	The department that checks the results	The first real users
Data	Example data only; nothing is stored	Copies of your data on a development system	Real data of real users
What you have at the end	A decision, the test findings and a first list of requirements	A measured result against the pass mark	A product in use, with numbers on where users stop
What it cannot tell you	Whether the technology holds under real use	Whether users want the product around the AI	How users beyond the first group will react

The other two have their own pages: AI proof of concept and MVP development.

Who I am

I am Michael Wutzke from Frankfurt. For more than twenty years I have planned and built online shops and customer portals, as an employee and as a consultant. At Policen Direkt I programmed the front end of a portal in which customers upload and manage their insurance policies, and interaction design was part of the job. Today I build with AI coding agents for a client’s platform. Details: Career stages.