Hooray, we have a shared idea, but is it actually any good?

The great moment at the end of an exhausting design studio workshop

Anyone who has ever planned and run a Design Studio workshop knows the moment at the end when everyone is completely exhausted and happy. Hours or even days of joint effort have turned the participants into a team that has developed a shared idea and is rightly proud of it. Applied professionally and facilitated with commitment, the method manages to bring every participant’s ideas onto a common denominator over several rounds.

That is where the strength of Design Studio lies. A workshop creates motivation and commitment; instead of persuasion, participation produces buy-in. The big advantage: whoever took part has contributed, knows how the idea came about, knows the idea itself and no longer needs to be convinced. Design Studio is excellent for developing shared ideas and visions and for generating buy-in among colleagues in very different roles and across all hierarchies — but the method itself does not provide for a reality check.

A workshop answers the question: can we agree on this direction? A reality check answers a different question: what happens when people who were not part of this shared creative process are confronted with the solution? This distinction is the heart of the matter. Design Studio workshops create ownership. But they do not automatically create evidence.

Consensus matters for organizations, but how good is the quality of the solution?

When teams talk about “the best idea” after a workshop, a different question is worth asking: best idea in what sense?

A look at research on idea generation shows that idea quality is multidimensional. Dean, Hender, Rodgers and Santanen examined evaluation models for ideas and describe four dimensions: “novelty, workability, relevance, and specificity” (Dean et al., 2006). An idea can therefore be new but not viable; it can be feasible but hardly relevant; it can be relevant but remain too vague. It is precisely this distinction that matters for Design Studio workshops.

The typical workshop often produces strong signals of resonance: the group understands the idea, can retell it and finds it plausible. But resonance is only one dimension. An idea can be attractive inside the workshop and still fail outside it — because of missing specificity, false assumptions or too much everyday friction.

There is a second problem on top of this: selecting creative ideas is not trivial either. In their open-access study on creative idea selection, Zhu, Ritter and Dijksterhuis summarise that people tend to “select mainstream ideas at the expense of creative ideas” (Zhu et al., 2021). This does not mean that every consensus-friendly idea is bad. It means: the dynamic that produces agreement does not automatically favour the boldest, the most relevant or the ultimately most successful solution.

This is particularly relevant for workshops, because they do not only generate ideas but often also condense and prioritise them. The phase in which many possibilities become one shared direction is socially highly charged. People want to be constructive. They do not want to look like blockers. They look for a solution that seems feasible within the given constraints. As a result, a group can agree on something that sounds very good but has met too little resistance from reality.

So the problem is not the workshop. The problem is when the workshop is mistaken for reality.

What gets left out

In hindsight, many technological products look as though they had been plausible long before they were genuinely fit for everyday use. Google Glass showed early on that technical possibility and social acceptance are not the same thing. Metaverse visions sounded big in strategy presentations but ran into banal and therefore decisive questions: who regularly puts on a headset for this? For which situation is the benefit so strong that the friction is accepted? With spatial computing products such as the Apple Vision, too, the central question is not only whether the technology impresses, but whether it fits human ways of working, routines, social situations and willingness to pay.

Such examples should not simply be dismissed as failures. They contain enormous technical achievement. That is precisely what makes them interesting. They show that a product cannot be measured by feasibility alone. A product has to fit people’s lives and their real needs.

A workshop can only answer this question to a limited extent. Participants can say that they find a concept helpful. They can find a journey plausible. They can defend an idea because they were involved in creating it. But in that moment they are not using the solution under real conditions. They are not paying for it. They are not integrating it into their working day. They do not have to live with the social side effects.

That is why the best shared idea needs a step that confronts it with reality.

Can prototypes and early user tests help with the reality check?

A prototype asks a different question than a workshop idea sketch. An idea sketch asks: can you imagine that this idea makes sense? A prototype asks: what do you do when you encounter it?

This shift is decisive. In a conceptual conversation, people react to descriptions. In a prototype test, they react to friction. They click in the wrong place. They overlook something. They interpret terms differently. They give up. They look for information the concept never anticipated. They use a feature differently than planned. Or they immediately understand something that still seemed complicated in the workshop.

Camburn et al. define a prototype as a “pre-production representation of some aspect of a concept or final design” (Camburn et al., 2017). That is a helpful, sober definition: a prototype is not the finished solution. It is a preliminary form that makes one particular aspect testable.

That is exactly where its value lies. A good prototype does not have to do everything. It has to make the riskiest assumption visible. If the greatest uncertainty is whether users understand the navigation, a clickable dummy may be enough. If the greatest uncertainty lies in a social situation, a scenario or a Wizard-of-Oz test is needed. If the greatest uncertainty lies in physical behaviour, a screen prototype will not do.

Research on low-fidelity prototypes also supports this nuanced view. In two experiments, Virzi, Sokolov and Karis found that low- and high-fidelity prototypes revealed “substantially the same sets of usability problems” (Virzi et al., 1996). That is a strong argument against the notion that a reality check is only possible with an almost finished product.

But the same source also urges caution. Even a high-fidelity prototype can be a “noisy approximation of the actual way” people use a product (Virzi et al., 1996). A test is therefore not an oracle. It is a controlled encounter with reality, not reality itself.

The reality check is possible and sound, but limited

This makes the argument more precise: early user tests with prototypes can deliver a sound reality check if it is clear which question they are meant to answer. They are particularly strong on comprehension, orientation, usability, perceived relevance and visible friction. They are weaker on willingness to pay, long-term use, habit formation and cultural acceptance.

That is not a weakness as long as you are aware of it. It only becomes dangerous when teams draw conclusions from a small test that are far too large. A prototype test can show that people misunderstand an idea. It can show that a central task does not work. It can show that a supposed benefit does not land. But it does not prove that a product will be successful.

The reality check after a workshop should therefore not ask: does this idea work overall? It should ask: which assumption is so risky that we have to test it before development?

This assumption might be:

• Users understand the benefit within the first minute.
• A task can be completed without explanation.
• The solution fits into an existing work situation.
• The social friction is acceptable.
• The perceived benefit is stronger than the effort.

Only once the assumption is clear does the prototype make sense. Otherwise you are just building a prettier argument.

Can AI help?

In the past, the step from workshop to testable artefact was often expensive and lengthy. After the workshop there were sketches, photos, a board, perhaps a concept deck and a lot of energy. Turning that into a realistic prototype took time, design capacity, development, content, data and coordination. That is why the reality check often stayed abstract. Preferences were tested instead of behaviour. Slides were shown instead of situations.

AI is changing this transition. It makes it easier to turn workshop results quickly into something testable: clickable interfaces, realistic sample content, simulated data, chat flows, service journeys, role simulations, simple tools or variants of a concept. Camburn et al. also describe prototyping techniques as a means of “lowering of cost and time” (Camburn et al., 2017). This is precisely where the practical relevance of AI lies: it lowers the threshold for bringing an idea into contact with users earlier.

That does not mean AI replaces reality. On the contrary: if AI is only used to produce even more convincing concept images internally, it reinforces the old problem. It then does not make ideas more testable, only more polished.

The more sensible use lies in testing faster. AI can turn the best workshop idea into a hypothesis, the hypothesis into a prototype, the prototype into a test scenario and observations into a new iteration. The sentence after the workshop is then not: this is the solution. It is: this is our best assumption. Now let us build it so that users can contradict it.

The state of research on AI in the product development process supports this direction, but with caution. In their open-access review, Witkowski and Wodecki show that there is still an “underutilization of AI in concept testing”. At the same time they see fewer studies on “product testing, validation, and post-launch optimization” (Witkowski & Wodecki, 2025). For this article that matters: the claim that AI accelerates prototype building is plausible and well supported. The claim that AI-supported prototypes already reliably predict market success would be too strong.

From buy-in to a testable assumption

The most important step after a Design Studio workshop is therefore not immediate implementation, but translation.

• “We are building a personal onboarding” becomes: new users understand without explanation what they should do next.
• “We are creating a new collaboration platform” becomes: people are willing to leave their existing routine because the new process is faster or clearer.
• “We are developing an immersive experience” becomes: the benefit is strong enough that people accept the physical and social friction of the medium.

Sentences like these are less inspiring than workshop posters. But they are more valuable, because they can be challenged. And only ideas that can be challenged make learning possible.

A good AI-supported reality check therefore consists of five steps:

• The best workshop idea is formulated not as a solution but as a hypothesis.
• The riskiest assumption is named explicitly.
• The prototype is built only as realistically as that assumption requires.
• Real users interact with the prototype on a task that is as relevant as possible.
• The team decides not by preference but by observed friction.

This preserves the value of the workshop without overstating it. Participation creates energy. The prototype brings resistance. Both are necessary.

Do not treat the first reality check as a bonus

Design Studio workshops are powerful because they get people thinking. They make perspectives visible, create a shared language and build ownership. In organisations that is neither soft nor incidental. Without buy-in, many good ideas die before they get a chance.

But buy-in is not the same as solution quality. An idea can be developed together, explained well and broadly supported internally and still fail in actual use. Not because the workshop was bad, but because workshops are social spaces. They produce social signals. Use only emerges outside.

AI makes the next step easier. It can help to get from a shared concept to a testable prototype faster. But it proves nothing on its own. The reality check only happens when real people meet a sufficiently concrete artefact in a sufficiently real task.

The best formula for modern product development is therefore not: more workshops or more AI. It is: a better connection between the two.

Design Studio workshops create ownership. AI-supported prototypes can make contact with reality possible earlier. Good product development needs both: participation and acceptance in the workshop, and contradiction from the world.