The question isn’t new, but artificial intelligence has given it a new sense of urgency. When we can achieve something in minutes that once took days of work, it’s natural to wonder whether the process still matters as much.
For a designer or developer, the process can be an essential part of the work. It’s where we understand the problem, explore alternatives and make decisions that aren’t always visible in the finished product. For someone managing projects or running a business, the outcome tends to carry more weight: what was delivered, how much it cost, when it was ready and what value it created.
These perspectives aren’t necessarily incompatible. They reflect different responsibilities for the same work. A designer needs to deliver and a manager needs to understand whether what has been delivered will hold up over time.
The problem begins when we turn either perspective into an absolute truth. Ah, absolute truths. They’ve never been the best guides.
A good process needs to serve a purpose
It’s possible to follow an exemplary process and end up with something that isn’t particularly useful. We can spend weeks researching, meeting, documenting and testing without addressing what the client actually needed. Rigour alone doesn’t guarantee relevance.
It’s also possible to reach a convincing result quickly through a weak process. An interface can look well designed while hiding usability problems. A piece of software can work in a demo and fail when it encounters unexpected data or more people start using it.
So perhaps the choice isn’t between process and outcome, but between what we need to deliver now and what it takes to make that work valuable, both now and later.
That doesn’t always mean finding a middle ground. Sometimes speed should take priority; at other times, certain steps simply cannot be skipped. Balance doesn’t have to mean an equal split. It has to fit the context.
AI makes outcomes more immediate and the process less visible
With artificial intelligence, a significant part of the work can happen without us following every step. We ask for a proposal, a design or a coding solution, and receive something that already looks like a developed product.
That capability is useful. It allows us to try more ideas, reduce repetitive tasks and bring proposals to life that might otherwise never get off the ground.
But it also changes how we perceive quality. When something looks finished, it’s easy to assume the problem has been solved. Presentation takes precedence over understanding, because what we see is immediate, while what we need to verify takes time.
Perhaps, then, it goes too far to say that AI has made the process irrelevant. What may be happening is that the process has become less visible. And a less visible process is easier to ignore.
The responsibility of anyone working with these tools now includes other questions: how do we know the solution is correct? What assumptions does it contain? What happens when it fails? Can we change it and maintain it?
The effort may shift from execution to defining the problem, assessing proposals and validating the outcome. Working faster doesn’t remove the need to make sound decisions.
Context determines how much rigour is needed
An image used to explore an idea, a prototype for a meeting and a system supporting essential operations don’t require the same level of oversight.
For work that carries little risk, is easy to correct and has a short lifespan, a lightweight process may be perfectly appropriate. Insisting on a lengthy one can therefore consume resources without adding enough value to justify them.
But when an error could compromise safety, cause significant losses or interrupt an essential service, the requirements change. In those cases, understanding how the solution was reached, testing its behaviour and establishing who is responsible are all part of the quality of what is delivered.
Many situations fall between these extremes. The decision depends on the deadline, the budget, the team’s availability, the client’s expectations and how long the solution will need to keep working.
A team juggling several projects may need to simplify. A client may accept a limited first version to test an assumption. A temporary solution may justify choices that would be inappropriate for a product intended to last for years. What matters is that these choices are deliberate and that the people making them understand the trade-offs. Above all, those trade-offs need to be clear to everyone.
Delivering quickly also means thinking about what comes next
Initial speed is only part of the equation. Time saved today can become time spent tomorrow if no one understands the solution, if every change introduces problems or if maintaining it means redoing the work. On the other hand, demanding every possible assurance upfront also comes at a cost. We may delay decisions, exhaust the budget or put enormous care into developing something no one needed. Perhaps the more useful question is this: how much process do we need to ensure a suitable outcome, at an acceptable level of risk, within the constraints we face?
AI makes this question more important because it makes it easier to produce answers before we have properly understood the questions. We don’t need to demonise speed or turn the process into an untouchable ritual. We need to know where to simplify, where to verify and where to stop and think.The outcome gives the work its purpose and the process helps us understand how much we can trust it.
Hélder da Rocha Pereira, Founder and Creative Director
