The conversation about AI replacing designers assumes the work in the near future will be done faster and efficiently just by swapping out who does it.

But that's not what's happening. The work itself is changing. And the designers who figure out where to let AI take over and where to stay in control are delivering better work faster than anyone thought possible several months ago.

It all comes to understanding which parts of the design process actually need a human making judgment calls and which parts are just... slow.

In this blog post, our lead product designer Nikola Dadić shares exactly how he used AI to compress the timeline, where it worked, where it didn't, and what it means for how we're thinking about product design going forward.

Why speed matters in today’s product development?

Great products take time to build. But the reality today is different because companies can't afford to spend years developing a new product when what's trending today might be irrelevant for users tomorrow.

The longer your development cycle runs, the wider the gap becomes between what you started building and what people actually need by the time you finish. 

A feature that made perfect sense in January can feel obsolete by July because the market conditions and user requirements changed.

The market rewards whoever ships a product first. Companies can leverage speed to become the name people associate with the solution. Everyone else spends their energy explaining how they're different, which is a harder position to be in.

The only way to close that gap is to compress how long it takes to go from idea to working prototype. Which means rethinking which parts of the process actually need weeks and which parts just feel like they should because that's how it's always been done.

The benefits of AI in product design

We can all agree that the traditional way of building digital products used to take some time. 

You had to move through multiple phases and involve several people before anyone could use a product the way it was intended. 

Today you can get something tangible and realistic in hours or days thanks to AI.

What AI enables for everyone, especially designers, product managers, and even clients, is the ability to build prototypes without technical knowledge. 

You don't need to know how to code a button or set up a development environment. The tools handle the technical execution while you focus on what the product should do and how it should feel.

When a client says they want a dashboard, you can show them a working version the next day instead of debating what "dashboard" means across several meetings. 

One clear benefit of this is that misunderstandings can get caught early, before they turn into expensive rebuilds.

How we use AI in the product design process (examples)

With multiple AI projects delivered this year, we refined a simple process that consistently accelerated our workflow and allowed us to deliver working prototypes to clients in days.

In this article we won’t be talking about a specific case study that will show how our process led us to deliver a certain solution for a client. We will cover the 4 simple steps that allowed us to speed up our product development process and deliver working prototypes in days.

1. Client meeting recording and transcript

The first client meeting captures how people actually describe their problem, what they care about, and what they expect to see working.

To make sure the design process stays connected to clients requirements, we record and transcribe the entire meeting.

Later on, that transcript becomes the input for AI tools like Chat GPT or Claude. Basically, we give AI clear instructions and correct it along the way to pull out key requirements, identify priorities, and flag contradictions that might not have been obvious in the meeting.

When design decisions stay tied to the client's actual words and stated priorities, there's less drift between the initial conversation and the prototype you build.

2. Create a Product Requirements Document (PRD)

Once the transcript is done, the next step is turning it into a PRD. 

The transcript gets passed to AI with a specific brief: describe a UX prototype using mock data. This keeps the focus on interaction and structure instead of getting stuck on backend logic or data sources that don't exist yet.

For better results the document needs structure. It must include user stories, user flows, and screens with their elements. 

Once the document is generated, the product team steps in. They review the output, identify gaps, and apply corrections. Only after that validation does the process move forward.

3. The implementation plan

Once the requirements are agreed on, they go straight back into AI to shape the implementation plan. In our case that happens in Claude Code, where the brief is simple and very practical. Turn the requirements into a clear path of execution.

Claude breaks the work down into steps it can follow, and that output gets reviewed by the team. 

The designer’s role here is to sanity-check direction, to make sure the direction and scope match what was actually discussed with the client.

4. Building a prototype with AI

In the last stage, there are a few hours of iteration process left where we check how elements work and feel together while making notes on what we need to change. 

Adjustments are applied as you go, based on what actually works in the interactions. That loop of reviewing and refining is where the quality of the prototype really comes together.

4 key insights on AI’s impact on product design workflow

1. Clear alignment at the beginning of the project

One of the biggest advantages of working with AI shows up long before design decisions. Early prototypes make expectations visible, which helps align clients and teammates while everything is still flexible.

When everyone reacts to the same thing on screen, assumptions surface fast and misunderstandings disappear early.

Seeing a working prototype makes it much easier to judge scope and answer the question everyone asks sooner or later which is: “How long will this actually take to build.”

2. Better time awareness

AI compresses the distance between idea and execution. Instead of guessing timelines based on abstract descriptions, teams can base estimates on something tangible. 

That leads to better planning and fewer surprises once development starts.

This matters especially in early phases, where wrong estimates tend to snowball. Faster feedback loops help keep expectations realistic on all sides.

3. Stronger design foundations

From a design perspective, AI helps establish a much stronger base. Early outputs give structure to ideas that would otherwise stay vague, which makes further design work more focused and intentional.

Clearer structure leads to better prototypes. Better prototypes create a more accurate sense of the final experience, not just how it looks but how it feels to use.

4. Faster validation with real users

When prototypes are more realistic, validation becomes easier and more meaningful. Clients can react to concrete interactions, and users can test flows that resemble the real product.

That shortens the path between idea and validation. Instead of debating concepts, teams observe behavior and adjust accordingly. 

That feedback is what ultimately protects the product from drifting away from real needs.

Conclusion

Most product requirements can easily lose its precision once they leave the meeting room.

The further it moves from the original conversation, the more it gets interpreted by people who weren't there, and the version of the product that eventually gets built reflects those interpretations as much as it reflects the original intent.

People are genuinely bad at knowing what they want in the abstract. The kind of clarity that produces real direction tends to arrive only in reaction to something tangible. 

A client who has spent days articulating a vision will sit down with a working prototype for the first time and find, within minutes, a more precise and more honest version of what they were trying to say all along. 

The thinking that was only half-formed in the room finishes itself the moment there is something real to push against.

Getting to that first moment of real reaction used to consume most of the project timeline. 

The team made its consequential decisions before anyone had touched anything, and the cost of being wrong about direction revealed itself only once the work was deep enough that changing it was genuinely painful.

That cost is what has changed. A designer who can put something real in front of a client on day two is working with a fundamentally different kind of information than one who can only do it in week four. 

The instinct that develops from working this way — the ability to read a room, catch a wrong direction early, and know when the product is drifting from what the client actually needs — is the kind that takes years to build through the old process.

The teams developing it now will carry that advantage into every project they take on from here.