Designing with AI now requires sound judgement more than ever
Artificial intelligence can speed up analysis, design, documentation and automation, but the real value still lies in knowing which problem to solve, how to use the tool, how to evaluate its results, and which conclusions are worth turning into a product.
Artificial intelligence · Product design · UX · Judgement · Innovation · Automation · Digital product
AI speeds up the work, but it does not replace management
AI can help you make faster progress, but it doesn’t decide for you which direction is worth pursuing.
In a very short space of time, artificial intelligence has become an everyday tool for researching, summarising, drafting, analysing, generating ideas, automating tasks and exploring alternatives. Many tasks that used to take hours can now be completed in minutes. This is a huge advantage, especially in jobs where you have to process large amounts of information, produce documentation, compare options or iterate on possible solutions.
But precisely because AI speeds up work so much, sound judgement becomes even more important. If a tool allows us to produce output more quickly, it also allows us to make mistakes more quickly, repeat superficial ideas more quickly or produce solutions that appear correct without having been sufficiently thought through. Speed is no substitute for direction.
The real difference will not lie in using AI to produce more, but in knowing how to guide it, question it and build something better based on its results. To stand out, create or innovate, it is not enough simply to ask a tool to generate options. You need to know what to ask for, how to interpret what it returns, what to discard, what to improve and what valuable conclusion can be drawn.
From using AI to thinking with AI
Using AI should not simply be about delegating tasks. It should serve to expand our analytical capabilities, speed up explorations and free up time to think more effectively. The difference lies in whether we use it as a substitute for judgement or as a tool that helps us work in greater depth.
AI can be very useful for automating repetitive tasks, organising scattered information, summarising documents, generating first drafts, detecting patterns, proposing alternatives or turning ideas into more polished materials. But that does not mean the result is automatically good, original, correct or useful.
Often, it will serve as a starting point, not a final answer. One of the current risks is confusing output with progress. Having more texts, more screens, more ideas or more options does not necessarily mean having a better product. AI can multiply possibilities, but without clear direction it can also multiply noise. It can give a sense of progress when in reality we are merely generating more material that will subsequently need to be interpreted, reviewed and prioritised.
There is also a creative risk. If many people use similar tools, with similar instructions and without a clear intention, the results tend to be similar. AI can help with creation, but the difference will still lie in the perspective, the context, the experience, the ability to connect ideas, and the judgement required to transform a generic output into something unique.
That is why working with AI requires an active approach. It is not a matter of simply accepting what it produces, but of steering the process. Asking better questions, providing context, setting constraints, cross-checking results, detecting biases, adjusting the tone, reviewing the logic, checking whether it suits the user, and deciding which parts are worth turning into a real solution.
The challenge lies in deciding what to delegate
AI can take on tasks, speed up processes and open up new possibilities, but not everything should be delegated in the same way. Designing with AI involves deciding what can be automated, what needs to be reviewed, what requires human supervision, and what continues to depend on professional judgement, experience and responsibility.
Delegate the task, not the responsibility
AI can generate a proposal, a summary, a recommendation or an automation, but the responsibility for its use remains with humans. Someone has to check whether the result is correct, whether it fits the context, whether it addresses the problem, and whether it can be used without causing harm, confusion or misguided decisions.
Speed with direction
Speed only adds value when there is a clear direction. If we do not know what problem we want to solve, what hypothesis we want to validate or what decision we want to support, AI can produce a great deal without helping us to make progress. The tool speeds things up, but it is our judgement that sets the course.
Creativity on the way out
Innovation is not about accepting the first answer generated, but about using it as raw material. Creativity emerges when we combine, question, reinterpret, discard and take what AI proposes a step further. The real value lies in what a person does with the output, not just in the output itself.
Designed supervision
Review must not be left as an ad hoc task at the end of the process. In products that incorporate AI, it is necessary to design the process by which a person validates, corrects, approves or blocks an output. Supervision is also part of the experience and must be integrated into the workflow.
Product criteria in the age of AI
Designing with AI is not about making a product appear more intelligent, but about deciding where artificial intelligence adds real value within a process.
In digital products, AI can help to conduct better research, explore alternatives, generate documentation, analyse data or prototype more quickly. But none of these capabilities replaces the need to understand users, the business, processes, technical constraints and risks. A solution may be very well generated and yet still fail to resolve anything of significance.
When AI is integrated into a product, this criterion becomes even more relevant. It is not enough to say that the system summarises, recommends, classifies or responds. We must decide what data it uses, how reliable the output is, how it is explained to the user, what can be corrected, which decisions are automated and what boundaries should not be crossed.
In sensitive sectors such as healthcare, public administration, digital identity, banking or data management, this consideration is even more important. AI can save time, detect patterns or facilitate decision-making, but it can also introduce errors, biases or false certainties if presented with excessive confidence. The more critical the context, the greater the need for an experience that makes human control, explanation and traceability visible.
The role of those of us who design is also changing. It is no longer enough to design static screens or closed workflows. We must design systems where some of the content, recommendations or decisions can vary depending on data, models or instructions. This requires us to consider uncertain states, imperfect results, human review, exceptions and trust.
To stand out in this context, creativity does not disappear; it shifts. It is no longer just about producing an idea from scratch, but about knowing how to formulate the problem effectively, select good references, guide the exploration, recognise useful patterns, connect different domains and turn a generated output into a purposeful proposal.
AI can reduce repetitive tasks and open up new possibilities, but it does not replace strategic thinking. If everyone has access to similar tools, the difference will lie in judgement: in the ability to know what deserves to be automated, what must be carefully designed, and what requires a human decision before becoming a product.
Thinking more effectively, not just producing faster
The future will not belong to those who use AI to produce more, but to those who know how to use it to think more clearly, make better decisions and build more useful products.
Artificial intelligence can be an extraordinary tool for speeding up work, reducing friction and expanding creative possibilities. But it should not become a shortcut that allows us to stop thinking. The easier it is to generate a response, the more important it is to know whether that response makes sense, whether it is well-targeted, whether it adds value and whether it deserves a place in the product.
Now more than ever, we need sound judgement. Judgement to ask questions, to review, to discard, to improve, to innovate and to take responsibility for what we build with the help of AI. The tool may speed up the process, but the direction still depends on our ability to understand the problem, provide context and turn possibilities into solutions with real value.
Let’s talk about AI, product and criteria
I work on digital products where artificial intelligence needs to be integrated in a meaningful way, bringing together technology, user experience and real business needs. If you’re exploring how to incorporate AI without losing sight of your goals, let’s have a chat.