Prototype

What is a prototype?

A prototype is a working model that shows how a product could look, feel or behave, without the effort of building it properly. It can be a set of clickable screens, a rough script that proves an integration is possible, or a quick app that chains a few AI calls together to show what an assistant might do.

It usually works in two steps. The team picks one question it needs answered, such as whether users can find the approval button or whether a model can read the company's invoices, and builds the smallest thing that answers it. Then it puts the prototype in front of users or stakeholders and records what happens.

Prototypes show up in product discovery, in design reviews, in sales conversations where a buyer needs to see an idea, and in technical spikes before a larger build. They are meant to be thrown away or rebuilt once they have done their job.

How is a prototype different from an MVP?

A prototype tests an idea; a minimum viable product (MVP) tests a business. A prototype can fake data, skip security and run only on a laptop, because its audience is a handful of testers or decision-makers. An MVP is used by real customers on real data, so it needs basic reliability, access control and support.

The difference matters for planning. Teams that treat a prototype as an MVP risk shipping something fragile. Teams that treat an MVP like a prototype spend too long polishing before they learn anything.

A common sequence is several quick prototypes to settle the design and the technical approach, followed by one MVP that turns the winning version into a product people can depend on.

What types of prototypes are there?

Prototypes range from sketches to working code, and each type answers a different question.

• Paper or wireframe prototypes: sketches and grey boxes to agree on structure and flow.
• Clickable design prototypes: linked screens in a design tool such as Figma, used for usability testing.
• Technical prototypes: small pieces of code that prove a hard part works, such as an integration or a data pipeline.
• AI prototypes: a quick test of whether a model can handle a task on the business's own examples, before anyone designs an interface around it.

The right type depends on the biggest open question. If the risk is whether users understand the product, test the design. If the risk is whether it can be built at all, test the technology first.

How fast can you build a prototype?

Often in days rather than weeks. Design tools, component libraries and AI coding assistants mean a clickable prototype or a basic working app can be assembled quickly. Approaches such as vibe coding, where a person describes what they want and an AI tool writes much of the code, suit this stage well.

That speed has a catch. A prototype that looks finished creates pressure to launch it as it is. Code generated for a demo rarely has the structure, tests or security needed for real users, so teams should agree upfront that the prototype is for learning, and plan a proper build if the idea survives.

The faster prototypes get, the more useful it is to test several options side by side instead of committing to the first one.

What makes a prototype test useful?

A useful prototype test has a clear question, the right testers and a record of what happened. Write the question down first, then choose people who match the intended users rather than colleagues who already know the idea.

During the session, give testers a realistic task and watch rather than explain. Where they hesitate, misread a label or work around a step, the design needs work. For technical prototypes, record what failed and under which conditions, not only the happy path.

Afterwards, summarize the findings as decisions: what to keep, what to change and what to drop. A prototype that does not change a decision was either testing the wrong question or was never really tested.

How does Vizio AI use prototypes?

In our Venture Studio, prototypes are how product strategy and UX ideas get tested before full software development begins. During idea validation we build clickable flows to check that users understand the product, and quick technical prototypes to confirm that an AI integration or data connection will work on real inputs.

When a prototype answers its question, we decide with the team whether to move into MVP development, change direction or stop. We keep prototypes and production code separate so that speed early on does not become a problem later. The usual starting point is a short discovery to pick the one question the first prototype needs to answer.