UX Design
AI and decision-making
Nov 16, 2026
AI makes design and prototyping faster, but more concepts create new challenges for review, feedback and decision-making. Explore how teams can adapt their design workflows.
AI Makes Prototyping Faster. But What About Design Reviews?
AI is changing how quickly teams can generate ideas, explore alternatives and create prototypes.
What previously took hours or days can now take minutes. Designers can explore more directions, stakeholders can generate their own concepts, and teams can produce more variations than ever before.
But there is a catch.
The ability to create more doesn't automatically increase our ability to evaluate, compare, review and make decisions.
In fact, as production becomes faster, another part of the design process starts to become increasingly visible: decision-making.
When creation becomes cheaper, what becomes more expensive?
Let's look at a typical feature development process.
A game designer comes up with a new feature. They write a specification, collect references and communicate the intended functionality to the team.
UX/UI designers explore possible solutions. They create wireframes, visual concepts and prototypes. Stakeholders review the work, provide feedback and request adjustments.
This process isn't new.
What's changing is the amount of work that can happen before the review stage.
With AI-assisted tools, a stakeholder can generate several interface concepts instead of describing one idea. A designer can explore multiple layouts in the time it previously took to create one. A developer can build a functional prototype to demonstrate a technical possibility.
This creates new opportunities for collaboration.
But it also increases the number of things that need to be evaluated.
More concepts mean more questions:
Which direction best supports the player's goal?
Which version fits the existing game experience?
Are we comparing equivalent solutions?
Which requirements are satisfied?
What are the implications for implementation?
Which feedback should influence the final direction?
The ability to generate alternatives is increasing. The time and attention needed to evaluate them haven't necessarily changed.
The new bottleneck: Review and decision-making
Traditionally, design production was often limited by the time required to create something.
Designers needed time to understand requirements, explore possibilities, prepare assets and build prototypes.
AI is reducing some of that effort.
But evaluating a solution still requires understanding the product, its players, its mechanics and its constraints.
A prototype can be generated in minutes. Understanding whether it is the right solution can take considerably longer.
This is particularly relevant in game development, where features rarely exist in isolation.
A new interface might affect progression, economy, navigation, notifications, player motivation and other interconnected systems.
A visually convincing prototype doesn't automatically account for all these relationships.
And when different stakeholders produce their own interpretations, the team can end up reviewing several disconnected concepts rather than making progress towards one shared direction.
AI can accelerate the production of alternatives. It doesn't automatically accelerate alignment.
Why traditional review processes struggle
Many teams still rely on a combination of specifications, screenshots, design files, comments, presentations and separate prototypes to review feature ideas.
Each format has its purpose, but together they can create fragmented conversations.
Consider a simple example.
A team is developing an Information Broker feature for a strategy game.
The feature owner imagines a full-screen intelligence experience. A UX designer proposes a character-driven interaction with an information panel. A developer creates a compact prototype that fits into the existing gameplay HUD.
All three concepts might be reasonable.
But reviewing them separately makes it harder to understand their differences and evaluate the trade-offs.
Are we discussing immersion? Information hierarchy? Screen space? Interaction complexity? Technical feasibility?
Without a shared context, feedback can become subjective and discussions can move between different assumptions.
The problem isn't that people have different ideas.
The problem is that they don't always have a shared environment in which to explore and compare them.
What if we moved decisions upstream?
Instead of treating review as a stage that happens after design production, what if we made collaborative evaluation part of the exploration process?
Imagine a workshop where stakeholders can work with the same interactive representation of the game.
The feature owner introduces the requirements and intended player outcome. UX provides the product context, established patterns and existing interactions.
Participants then explore different interpretations of the feature.
They can compare alternatives, challenge assumptions and discuss details while looking at the same product context.
Rather than exchanging disconnected screenshots and comments, the team works towards a shared understanding of the intended experience.
This is what I call Collaborative Experience Calibration.
Collaborative Experience Calibration
The idea is to bring stakeholders together to explore and align on feature details before those decisions become expensive to change.
A possible workflow:
1. Start with a shared baseline
Provide an interactive representation of the game, including relevant components, patterns, gameplay context and existing behaviours.
2. Introduce the feature
The feature owner explains the problem, player goals, requirements and constraints.
3. Explore different interpretations
Participants create alternative versions of the feature using the shared foundation.
4. Compare and challenge
Review the alternatives against player needs, interaction behaviour, product consistency and production implications.
5. Agree on the direction
Identify the strongest solution, resolve open questions and document decisions.
6. Integrate validated outcomes
Update the shared prototype and make the agreed direction available to the wider team.
The workshop doesn't need to resolve every implementation detail. Its purpose is to reduce ambiguity, align expectations and make important experience decisions earlier.
Where does the Design Harness fit in?
This is one of the reasons I started exploring the concept of a Design Harness.
A Design Harness provides the shared environment needed to make this kind of collaboration possible.
It connects product context, design foundations, UX patterns, interaction behaviour and interactive prototypes.
It gives stakeholders a common starting point instead of asking everyone to interpret disconnected references or build everything from scratch.
AI can become a valuable part of this environment.
It can help generate variations, populate prototypes with existing components, identify inconsistencies, compare alternatives and document workshop outcomes.
But AI isn't the foundation of the idea.
The foundation is shared context.
Without it, AI can simply help us produce more disconnected interpretations, faster.
The real opportunity isn't creating more. It's deciding better.
As AI continues to reduce the cost of creating concepts and prototypes, the ability to make good decisions becomes increasingly important.
Design teams will need to rethink not only how they produce interfaces, but also how they collaborate, review and establish direction.
We shouldn't measure progress only by how quickly a team can generate a prototype.
We should also consider how effectively it can evaluate that prototype, align stakeholders and turn the result into a coherent player experience.
When creating ideas takes minutes, spending days trying to agree on them becomes an increasingly expensive part of game development.